Serum cytokines predict efficacy and toxicity, but are not useful for disease monitoring in lung cancer treated with PD-(L)1 inhibitors

Abstract

Introduction:

PD-(L)1 inhibitors (IO) have improved the prognosis of non-small-cell lung cancer (NSCLC), but more reliable predictors of efficacy and immune-related adverse events (irAE) are urgently needed. Cytokines are important effector molecules of the immune system, whose potential clinical utility as biomarkers remains unclear.

Methods:

Serum samples from patients with advanced NSCLC receiving IO either alone in the first (1L, n=46) and subsequent lines (n=50), or combined with chemotherapy (ICT, n=108) were analyzed along with age-matched healthy controls (n=15) at baseline, after 1 and 4 therapy cycles, and at disease progression (PD). Patients were stratified in rapid progressors (RP, progression-free survival [PFS] <120 days), and long-term responders (LR, PFS >200 days). Cytometric bead arrays were used for high-throughput quantification of 20 cytokines and other promising serum markers based on extensive search of the current literature.

Results:

Untreated NSCLC patients had increased levels of various cytokines and chemokines, like IL-6, IL-8, IL-10, CCL5, G-CSF, ICAM-1, TNF-RI and VEGF (fold change [FC]=1.4-261, p=0.026-9x10^-7) compared to age-matched controls, many of which fell under ICT (FC=0.2-0.6, p=0.014-0.002), but not under IO monotherapy. Lower baseline levels of TNF-RI were associated with longer PFS (hazard ratio [HR]= 0.42-0.54; p=0.014-0.009) and overall survival (HR=0.28-0.34, p=0.004-0.001) after both ICT and IO monotherapy. Development of irAE was associated with higher baseline levels of several cytokines, in particular of IL-1β and angiogenin (FC=7-9, p=0.009-0.0002). In contrast, changes under treatment were very subtle, there were no serum correlates of radiologic PD, and no association between dynamic changes in cytokine concentrations and clinical outcome. No relationship was noted between the patients’ serologic CMV status and serum cytokine levels.

Conclusions:

Untreated NSCLC is characterized by increased blood levels of several pro-inflammatory and angiogenic effectors, which decrease under ICT. Baseline serum cytokine levels could be exploited for improved prediction of subsequent IO benefit (in particular TNF-RI) and development of irAE (e.g. IL-1β or angiogenin), but they are not suitable for longitudinal disease monitoring. The potential utility of IL-1/IL-1β inhibitors in the management and/or prevention of irAE in NSCLC warrants investigation.

Introduction

Non-small-cell lung cancer (NSCLC) is the deadliest malignancy with an estimated 1.8 million deaths worldwide in 2021 (1). Most patients are diagnosed with advanced, incurable disease and a median life expectancy below two years (2). Immunotherapy (IO) with programmed death-(ligand) 1 [PD-(L)1] inhibitors, like pembrolizumab, atezolizumab and nivolumab, was a major step forward in the management of stage IV disease, facilitating long-term disease control and 5-year overall survival (OS) rates of 20-30% (3, 4). However up to 1/3 of patients do not respond, while potentially life-threatening grade III-IV immune-related adverse events (irAE) occur in approximately 10% of cases (5–7). One major unmet need is finding more reliable predictors of efficacy and toxicity to improve guidance of patient management. Tumor PD-L1 is the only currently approved biomarker (8, 9), and has also demonstrated association with the development of oligoprogression under IO (10), as well as irAE (5), but these associations are weak and complicated by considerable spatial heterogeneity of PD-L1 expression (11). Besides, the tumor mutational burden (TMB) is a pure genetic biomarker less prone to sampling errors, whose implementation has nevertheless been hampered by insufficient predictive potential and considerable technical variability (12–14). Immunologic parameters of the tumor tissue, like the emerging association of B cells and tertiary lymphoid structures with long-term IO benefit (15, 16), are an attractive alternative that directly reflects immunobiologic processes (17), but their use is limited by the scant material obtainable through small biopsies and the high procedural risk of repeated assessments. Therefore, there is increasing interest in soluble blood biomarkers that could be used to stratify patients and monitor treatment in a non-invasive manner (18). While rudimentary parameters based on routine laboratory tests, like the neutrophil-to-lymphocyte ratio (NLR, from the differential hemogram), and the advanced lung inflammation index (ALI, also incorporating the body-mass-index and serum albumin), have demonstrated predictive and prognostic utility for IO-treated NSCLC (19), a more detailed analysis of immunologic effector molecules in the blood would be expected to provide refined information and thus improve accuracy (17). Aim of this study was to systematically investigate the potential clinical utility of serum cytokines for the management of NSCLC patients treated with PD-(L)1 inhibitors.

Materials and methods

Patients and samples

This study included all patients with metastatic NSCLC and available serum samples, who received immunotherapy in the Thoraxklinik Heidelberg between 2012-2020, with a data cut-off on October 12^th 2021. PD-(L)1 inhibitors were administered either alone in the first (1L-IO) or subsequent lines (2⁺L-IO, after preceding chemotherapy), or in the 1L combined with chemotherapy (ICT). Serum samples were collected prospectively at baseline before treatment start, and longitudinally after 1 (1C) and 4 cycles (4C) of therapy, as well as at the time of disease progression (PD) (20). In order to capture more clear signals of efficacy, the focus was placed on cases with either rapid progression (RP), i.e. progression-free survival (PFS) < 120 days, or long-time response (LR), i.e. PFS > 200 days, while patients with intermediate PFS (120-200 days) were excluded from analysis. A group of age-matched healthy subjects without NSCLC was analyzed as controls.

Histological diagnosis and molecular profiling of NSCLC using combined DNA/RNA next-generation sequencing (NGS) were performed in the Institute of Pathology Heidelberg, as published (21). Patients with routinely treatable genetic alterations, like EGFR and BRAF-V600 mutations, or ALK/ROS1/RET/NTRK fusions, received tyrosine kinase inhibitors and were excluded from this study. The only two cases with mutations of these genes were one patient with BRAF p.G466E, for which no targeted therapy has been approved yet by the EMA or FDA, and one case with MET exon 14 skipping, for which MET inhibitors had not been approved yet at the time of the patient’s treatment, both of which received first-line chemoimmunotherapy. Clinicopathological parameters were collected from the patients’ records. The following parameters were extracted: demographics, baseline clinical and tumor characteristics including the Eastern Cooperative Oncology Group performance status (ECOG PS) and smoking status, PD-L1 tumor proportion score (TPS), results of differential blood counts, irAE characteristics, systemic anticancer treatments, date of progression, date of the last follow-up, and date of death. PD-L1 TPS was assessed using the clone SP263 (Ventana/Roche, Mannheim, Germany) and trichotomized for analysis as <1, 1–49, and ≥50%. IgG and IgM against human cytomegalovirus (CMV) was quantified using ELISA according to the manufacturer’s instructions (Euroimmun, Lübeck, Germany), and presence of either antibody class was considered to reflect positive serologic status.

PFS was defined as the time from immunotherapy start to death or progression. Overall survival (OS) was defined as the time from immunotherapy start to death or last follow-up. The progression date under immunotherapy was verified by the investigators with review of radiologic images, i.e. chest/abdomen CT and brain MRI-based restaging every 6–12 weeks, without formal RECIST reevaluation, as several studies have demonstrated very good agreement between real-world and RECIST-based assessments (22, 23). Diagnosis of irAE was based on standard clinicolaboratory criteria (24). This study was approved by the ethics committee of Heidelberg University (S-579/2019), and all participants gave informed consent.

Selection of target cytokines

The cytokines in the panel were selected based on a search for original articles in PubMed on serum markers potentially associated with PD-(L)1 inhibitor efficacy in NSCLC or the development of irAE. Search terms were ((predictive biomarker[Title/Abstract]) AND ((NSCLC[Title/Abstract]) OR (lung cancer[Title/Abstract]))) with 624 results. Review articles, publications about treatments other than immunotherapy, about other tumor entities (e.g. SCLC), or about non-soluble biomarkers were excluded. After compilation of the first database, 10 additional papers were found through manual search focused on already identified potential markers. According to the published evidence and technical feasibility of multiplex measurements, the following 20 markers were selected for the current analysis: interleukin (IL)-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17F, interferon gamma (IFN-γ), tumor necrosis factor (TNF), intercellular adhesion molecule 1 (ICAM-1), interferon-gamma induced protein 10 (IP-10), vascular endothelial growth factor (VEGF), angiogenin, soluble CD40 ligand (sCD40L), granulocyte-colony stimulating factor (G-CSF), CC motif chemokine ligand 5 (CCL5), granzyme a, and soluble TNF-receptor I (TNF-RI). The results of the literature search and the rationale for selection of analyzed markers are shown in Supplementary Table 1.

Sample processing and cytokine testing

Blood was collected in lithium-heparin tubes, centrifuged at 2000 g for 10 minutes within 1 h from venipuncture, followed by removal, aliquoting and storage of serum at -80°C. For the quantification of cytokines, aliquots were thawed on ice and measured using cytometric bead arrays (CBA) according to the manufacturer’s instructions (Beckton Dickinson, Heidelberg, Germany) with standard (S) or enhanced (E) sensitivity kits, as appropriate. The limit of detection and range of assays used in this study is shown in Supplementary Table 2. In brief, each sample was centrifuged at 12000 g for 2 min at 4°C, and the supernatant transferred into a new tube and diluted 1:3 (S) or 1:4 (E) to a total volume of 50 µl. Capture beads were diluted 1:20 (for E kits, followed by a single wash step using 1 ml of wash buffer at 200g for 5 minutes) or 1:50 (for S kits), added to the samples (E: 20 µl; S: 50 µl), and incubated at room temperature (E: 2 h; S: 1 h). Next, the phycoerythrin (PE) detection reagent was added (E: 20 µl diluted 1:20; E: 50 µl diluted 1:50), samples were incubated for 2 h at room temperature in the dark, for E kits the second component of the detection diluent 1:10 diluted was added, and all samples underwent a final washing step with resuspension in 200 μl of wash buffer for measurement. Standard curves were generated by processing the lyophilized standards provided with the kits in a similar way as the patients’ samples. For sample acquisition, an LSR-Fortessa Flow Cytometer (Beckton Dickinson, Heidelberg, Germany) was used. Cytokine concentrations were calculated from the raw CBA data using the Fcap Array™ version 3.0 software (Soft Flow, Pecs, Hungary).

Statistical analysis

Statistical comparisons between patient groups (e.g. RP vs. LR) were performed using Wilcoxon tests, while paired Wilcoxon tests were used to analyze different time-points of the same patients (e.g. baseline vs. 4C). Fold change (FC) was calculated trough division of mean values. Survival was analyzed according to Kaplan-Meier and compared between groups using log-rank tests, after determining the optimal cut-off based on ROC and Youden-index analysis. The association with various parameters with survival was explored using Cox regression. The correlation between clinicopathological variables and serum markers was analyzed according to Spearman, while correlations classified as very weak (|r|<0.2), weak (|r|=0.2-0.3), moderate (|r|=0.3-0.5), and strong (|r|=0.5-0.7). Multiple testing correction was performed according to Benjamini- Hochberg. Statistical calculations were performed with SPSS version 28 (IBM Corp., Armonk, NY, USA) and R version 4.2.1 (www.R-project.org). Two-tailed p-values lower than 0.05 and with false discovery rate (FDR) lower 0.1 were considered significant.

Results

Patient characteristics

A total of 204 patients with metastatic NSCLC who received PD-(L)1 inhibitors in the first or subsequent lines could be included in the study (Figure 1). An overview of characteristics for study patients is given in Table 1, while more details about treatment type and stratification based on IO efficacy are provided in Supplementary Table 3. Mean age was 65 years (range 37-87) for first-line patients (n=154), 63 years (range 48-78) for patients in subsequent lines (n=50, Table 1), and 66 years (range 58-81) for age-matched healthy donors (n=15).

Figure 1

Serum cytokine profile of advanced NSCLC at baseline

Several cytokines were significantly increased in the serum of untreated patients compared to age-matched healthy controls, i.e. IL-6, IL-8, IL-10, CCL5, G-CSF, ICAM-1, TNF-RI and VEGF (fold-change [FC]=1.4-261, p=0.026-9x10^-7, Table 2 and Figure 2). Besides, chemotherapy-pretreated 2⁺L patients before start of immunotherapy in later lines showed similar changes, with significantly increased serum concentrations of IL-6, IL-8, IL-10, IP-10, CCL5, ICAM-1, TNF-RI and VEGF compared to the controls (FC= 1.7-2839; p=0.017–7x10^-8), but mixed changes compared to untreated, newly diagnosed patients: IL-6, IP-10, CCL5 and ICAM-1 were significantly increased (FC= 1.4-11; p=0.011–0.0007), while IL-2 and G-CSF were decreased (FC= 0.2-0.3; p=0.005–0.008, Table 2 and Figure 2). Of note, the baseline NLR was increased in all patient subgroups compared to controls, but decreased in 2⁺L compared to untreated patients (Table 2).

Figure 2

Serum cytokine changes under treatment with PD-(L)1 inhibitors

In contrast to prominent aberrations at the time of immunotherapy start, serum cytokine changes under treatment with PD-(L)1 inhibitors were subtle (Figures 3A, B). Most consistent was a decrease in angiogenin after 1 cycle in the first line (FC=0.6-0.7, p=0.014-0.006) or after 4 cycles in later lines (FC=0.3, p=10^-5). Other changes were either inconsistent, i.e. ICAM-1 dropped under IO-monotherapy in the first line, but increased under treatment in later lines; sCD40L was lower under treatment in the first, but not in subsequent lines; CCL5 was decreased under treatment in subsequent, but not in the first line; or concerned chemoimmunotherapy only, but not PD-(L)1 monotherapy, i.e. decreases of IL-6, IL-8, IL-12p70, and increases in IP-10 and G-CSF (Figures 3A, B). The NLR also decreased under treatment, but only in patients receiving ICT (Figures 3A, B). No significant changes were observed at the time of disease progression compared to the levels after 4 treatment cycles (Figures 3A, B).

Figure 3

Serum cytokine changes associated with survival

For both first-line PD-(L)1 inhibitor monotherapy and ICT cohorts, LR patients had significantly lower TNF-RI levels at baseline compared to RP patients (FC=0.5-0.8, p=0.006-0.003, Table 2). The respective TNF-RI cut-off was 2139.7 pg/ml, as determined by receiver operating characteristic (ROC) curve and Youden index analysis (Supplementary Figure 1). Patients with low TNF-RI baseline levels receiving PD-(L)1 monotherapy showed longer PFS (442 vs. 80 days in median, hazard ratio [HR] = 0.42, p=0.014) and OS (not reached vs. 229 days, HR = 0.28, p=0.004) compared to patients with high TNF-RI levels, Figures 4A, B). Besides, patients with low TNF-RI baseline levels receiving ICT showed longer PFS (409 vs. 212 days in median, hazard ratio [HR] = 0.53, p=0.009) and OS (not reached vs. 493 days, HR = 0.34, p=0.001) compared to patients with high TNF-RI levels, Figures 4C, D). Additionally, lower IL-6 levels were also significantly linked with IO efficacy for patients receiving ICT (Table 2), but other associations did not exceed the FDR<0.1 threshold (Supplementary Table 4). Patients with lower IL-6 levels at baseline receiving ICT showed longer PFS (436 vs. 212 days in median, HR = 0.50, p=0.003) and OS (not reached vs. 514 days in median, HR = 0.29, p=0.0003) than patients with higher IL-6 levels. No differences in cytokine levels under treatment and during disease progression were observed between LR and RP after correction for multiple testing (Table 2 and Supplementary Table 4). The NLR at baseline and after 4 cycles of treatment was associated with LR in patients receiving PD-(L)1 monotherapy, but not in patients receiving ICT (Table 2).

Figure 4

Baseline serum cytokines levels associated with the development of irAE

Several cytokine abnormalities were evident in baseline samples of patients receiving ICT, who subsequently developed irAE (Figures 4E, F), i.e. in order of decreasing degree of association: elevated baseline levels of angiogenin (FC=9, p=0.0002), IL-1β (FC=7, p=0.009), IL-5 (FC=4, p=0.011), IL-12p70 (FC=4, p=0.019), IL-10 (FC=3, p=0.004), and IL-8 (FC=1.6, p=0.0285), as well as reduced baseline levels of G-CSF (FC=0.7, p=0.007). In addition, patients developing irAE under PD-(L)1 monotherapy in later lines showed higher baseline levels of IL-5 (FC=17, p=0.003), IL-12p70 (FC=9, p=0.014), and granzyme A (FC=1.3, p=0.003, Table 2). The characteristics of irAE are shown in Supplementary Table 5. No significant changes of the aforementioned cytokines were noted according to the severity or irAE (grade 1/2 vs. 3/4) or the use of steroids or not (data not shown).

Serum cytokine levels associated with clinical characteristics

Several cytokines were associated with each other, most notably IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, ICAM-1, TNF and angiogenin formed a cluster with multiple (>1) moderate (r>0.3) significant correlations with each other (Table 3). Notable was also a weak, but significant (r=0.28, p=5x10^-9) correlation between TNF-RI and IL-6, the two cytokines more prominently linked to immunotherapy efficacy, as described in the previous section Serum cytokine changes associated with survival.

Furthermore, the correlations between baseline serum cytokine levels and clinical characteristics are summarized in Table 4. IL-6 levels were higher in men (FC=4, p=0.003), while IP-10 showed a moderate positive correlation with age (r=0.323, p=4x10^-5). All other associations noted were either non-significant or weak (|r|<0.3). In particular, no association was found between the serologic CMV status of patients and serum cytokine levels.

Discussion

Main objective of this study was to characterize the potential clinical utility of serum cytokine concentrations in IO-treated NSCLC. One prominent finding was the profoundly altered cytokine profile of newly diagnosed NSCLC patients compared to age-matched healthy controls (Table 2 and Figure 2). Several interrelated (Table 4) mediators were significantly elevated, including the proinflammatory cytokines IL-6, IL-8, G-CSF and the proinflammatory chemokine CCL5 (25), the proinflammatory adhesion glycoprotein ICAM-1, soluble TNF-RI, which is increased in inflammatory states to curb the bioactivity of TNF (26, 27), the anti-inflammatory IL-10 (28), as well as the proangiogenic VEGF (29). These results illustrate the systemic inflammation and immune dysregulation present in metastatic NSCLC, which explains several disease manifestations and offers specific therapeutic vulnerabilities. For example, the increased NLR in the blood of these patients is facilitated by elevated levels of various interleukins and G-CSF, which stimulate granulopoiesis (30), ICAM-1 expression is induced by inflammation and associated with worse prognosis in NSCLC and other cancers likely by facilitating the metastatic cascade (31), while the therapeutic relevance of elevated VEGF in the circulation is reflected by the success of bevacizumab and other angiogenesis inhibitors in the treatment of metastatic NSCLC (32). Of note, association of blood cytokine levels with the tumor PD-L1 expression were generally absent or very weak (Table 4), so that these parameters capture different aspects of NSCLC immunobiology. For example, alterations of the lung and gut microbiome were associated with increased levels of several inflammatory serum cytokines, including IL-6 and TNF-α, in preclinical models of lung cancer, which could in part be remedied by the administration of probiotics (33, 34). At the same time, the association of IP-10 and CCL-5 with PD-L1 TPS (Table 3) probably explain why these were elevated only in 1L-IO (with higher PD-L1 expression ≥ 50%) or ICT patients (with lower average PD-L1 expression), respectively (Table 2).

In contrast, serum cytokine changes under treatment were subtle. Consistent was only a decrease in angiogenin, also known as ribonuclease 5, a small, 123 amino acid protein that stimulates angiogenesis alongside several other pleotropic effects (35). Recently it has been reported, that increased serum angiogenin correlated with dynamic contrast-enhanced MR-PET parameters in NSCLC patients, which were improved under anti-angiogenic therapy and linked with OS (36). However, in our study the change in angiogenin under immunotherapy was neither accompanied by a decrease in circulating VEGF nor associated with OS (Table 2), therefore its significance remains unclear. From a clinical perspective, an important conclusion from the results of this study is that serial serum cytokine measurements are not suitable for disease monitoring, since there were minimal changes under treatment in our cohort (Figures 3A, B). This is in contrast to longitudinal ctDNA assays, which have demonstrated potential clinical utility in the context of both immunotherapy-treated (37) and oncogene-driven disease (38), based on the strong association of higher tumor mutation levels in the blood under therapy with refractory disease and shorter PFS (39). In particular, serum cytokines are obviously not suitable for early detection of treatment failure, since no consistent changes accompanied PD in our patients, contrary to ctDNA-based liquid biopsies, which can reveal emergence of novel mutations of increases in allelic frequencies of preexisting variants as a sensitive marker of PD several months earlier than radiologic tumor growth according to recent pivotal studies (40). Moreover, dynamic changes of cytokine levels after 1 and 4 treatment cycles were not associated with immunotherapeutic efficacy (LR vs. RP, Table 2), while dynamic changes of the NLR after 4 cycles (12 weeks) correlated with clinical outcome, as has also been observed by other investigators (41). Besides low sensitivity, as demonstrated by the current study, another problem of disease monitoring using serum cytokines would be susceptibility to external influences by factors unrelated to the tumor remission status, like use of steroids and concomitant infection (42, 43). The lack of association between serum cytokine concentrations in NSCLC and the serologic CMV status (Table 4), itself linked to mild chronic immune activation and immunosenescence (44), also reflects the inability of cytokines to capture subtle systemic changes of the adaptive immunity, as those expected to occur longitudinally under PD-(L)1 blockade.

Another important question is whether baseline cytokine levels could be used for improved prediction of immunotherapy benefit. In general, few differences in the blood levels of analyzed cytokines between patients with LR vs. RP were noted, collectively suggesting an association between lower levels of inflammatory markers, such as IL-6, IL-8, IP-10, TNF-RI and the NRL, with better immunotherapy outcome (Table 2). Similar observations have recently been reported by other investigators, as well, for example lower IL-6 and IL-8 levels at baseline as well as after 1 cycle of treatment were strongly linked to longer survival under immunotherapy in patients with lung cancer and melanoma from a prospective multicenter study in Italy (45). Based on the results of the current study, low levels of TNF-RI appear to be a particularly promising marker for several reasons: first, for patients receiving PD-(L)1 inhibitor monotherapy, the TNF-RI differences between LR vs. RP were more pronounced than those observed for the established marker NLR (FC=0.5 with p=0.003 vs. FC=0.8 with p=0.024, Table 2) (46); second, TNF-RI retained prognostic utility also for patients treated with ICT (FC=0.8 with p=0.006, Table 2 and Figure 4), which is a major unmet need, because the NLR, ALI, PD-L1 and other biomarkers of PD-(L)1 monotherapy become useless, when additional chemotherapy is administered (19); third, blood TNF-RI levels showed no correlation with tissue PD-L1 expression (Table 4), which means that it represents an independent biomarker; finally, the signal of TNF-RI observed in this study appears to be stronger than that of several other inflammatory biomarkers described in the literature, like the blood levels of IL-6 (47), IL-8 (48), IP-10 (49), ICAM-1 (50) and VEGF (51). Actually, all these molecules showed significantly increased levels at baseline in our patients, similar to TNF-RI, but lower levels of TNF-RI could much better discriminate LR vs. RP patients (Table 2 and Supplementary Table 4).

Remarkable was also the association of several serum cytokines, in particular increased angiogenin and IL-1β, but also IL-5, IL-8, IL-10, IL-12p70, and granzyme A, or decreased G-CSF at baseline, with the subsequent development of irAE (Figures 4E, F). Higher baseline IL-1β, IL-10 and IL-12p70 serum concentrations in NSCLC patients who later developed rheumatic irAE have been independently confirmed in a different patient cohort (personal communication with KB and MMSC). Such an association between preexisting systemic inflammation and autoimmunity has also been observed in other tumor types, like indolent B-cell lymphomas (52) and malignant thymoma (53). These results corroborate previous reports of a higher propensity for the development of irAE in IO-treated cancer patients with higher blood IL-10 (54) and other inflammatory mediators (55). Of note, irAE-related cytokines did not include TNF-RI and IL-6, whose lower levels were associated with longer survival in this study (Table 2), so that a complex protein panel may be able to independently predict both IO efficacy and IO toxicity. The association of baseline IL-1β levels with irAE is particularly interesting, because it had not been reported in NSCLC patients before, and because the IL-1β inhibitor canakinumab and other IL-1 drugs are widely, in part off-label, used to treat a variety of mainly autoinflammatory disorders in rheumatology (56). There is also evidence suggesting anticancer activity of canakinumab, for example its use was associated with a reduced incidence of lung cancer in the phase 3 CANTOS study (57), so that phase 3 trials of this drug in combination with chemotherapy for the treatment of NSCLC in various stages are ongoing (58). The findings of this study suggest that these drugs could potentially be useful in the treatment or even prevention of irAE, as well.

Main advantages of this work are the relatively large number of cytokines based on a preceding systematic literature review, the relatively large number of patients in several dedicated cohorts, i.e. first-line PD-(L)1 monotherapy, ICT, or PD-(L)1 monotherapy in subsequent lines, the prospective longitudinal sample collection at defined uniform time-points, the simultaneous consideration of IO efficacy and toxicity, and the rigorous statistical testing including correction for multiple comparisons and multivariable testing. Main limitations are the smaller number of available samples under treatment and at disease progression, as well as the inability to exclude potential confounders. Therefore, the results will need to be validated in future studies, which might pave the way for building a complex score based on several cytokines and data mining analysis (59). Other emerging approaches to refine patient stratification are measurement of circulating tumor cells, ctDNA, miRNA, blood exosomes, gene expression profiling, or analysis of the T-cell receptor (TCR) repertoire (18, 37, 60–64).

Conclusion

In conclusion, this study could demonstrate that several altered serum cytokines in patients with advanced NSCLC could be exploited in order to predict efficacy and toxicity of PD-(L)1 monotherapy or ICT more accurately, but they are not suitable for longitudinal disease monitoring and early detection of tumor escape.

Funding

This work was supported by the German Center for Lung Research (DZL 3.0).

Acknowledgments

We would like to thank Simone Kuder for help with the data collection, Ingrid Heinzmann-Groth for help with the collection of patient samples, as well as Tamara Hedinger, Simone Karcher-Bausch and Simone Butz for expert technical assistance.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

• 1

  SungHFerlayJSiegelRLLaversanneMSoerjomataramIJemalAet al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin (2021) 71:209–49. doi: 10.3322/caac.21660

  • CrossRef
  • Google Scholar

• 2

  SiegelRLMillerKDFuchsHEJemalA. Cancer statistics, 2021. CA Cancer J Clin (2021) 71:7–33. doi: 10.3322/caac.21654

  • CrossRef
  • Google Scholar

• 3

  ReckMRodríguez-AbreuDRobinsonAGHuiRCsősziTFülöpAet al. Five-year outcomes with pembrolizumab versus chemotherapy for metastatic non-Small-Cell lung cancer with PD-L1 tumor proportion score ≥ 50. J Clin Oncol (2021) 39:2339–49. doi: 10.1200/JCO.21.00174

  • CrossRef
  • Google Scholar

• 4

  GrayJRodriguez-AbreuDPowellSHochmairMGadgeelSEstebanEet al. FP13.02 pembrolizumab + pemetrexed-platinum vs pemetrexed-platinum for metastatic NSCLC: 4-year follow-up from KEYNOTE-189. J Thorac Oncol (2021) 16:S224. doi: 10.1016/j.jtho.2021.01.141

  • CrossRef
  • Google Scholar

• 5

  DanielloLElshiatyMBozorgmehrFKuonJKazdalDSchindlerHet al. Therapeutic and prognostic implications of immune-related adverse events in advanced non-Small-Cell lung cancer. Front Oncol (2021) 11:703893. doi: 10.3389/fonc.2021.703893

  • CrossRef
  • Google Scholar

• 6

  KhojaLDayDWei-Wu ChenTSiuLLHansenAR. Tumour- and class-specific patterns of immune-related adverse events of immune checkpoint inhibitors: a systematic review. Ann Oncol (2017) 28:2377–85. doi: 10.1093/annonc/mdx286

  • CrossRef
  • Google Scholar

• 7

  SantiniFCRizviHPlodkowskiAJNiALacoutureMEGambarin-GelwanMet al. Safety and efficacy of re-treating with immunotherapy after immune-related adverse events in patients with NSCLC. Cancer Immunol Res (2018) 6:1093–9. doi: 10.1158/2326-6066.CIR-17-0755

  • CrossRef
  • Google Scholar

• 8

  PlanchardDPopatSKerrKNovelloSSmitEFFaivre-FinnCet al. Metastatic non-small cell lung cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol (2018) 29:iv192–237. doi: 10.1093/annonc/mdy275

  • CrossRef
  • Google Scholar

• 9

  PlanchardDPopatSKerrKNovelloSSmitEFFaivre-FinnCet al. ESMO clinical practice living guidelines - metastatic non-Small-Cell lung cancer (2021). Available at: https://www.esmo.org/guidelines/lung-and-chest-tumours/clinical-practice-living-guidelines-metastatic-non-small-cell-lung-cancer.

  • Google Scholar

• 10

  RheinheimerSHeusselC-PMayerPGaissmaierLBozorgmehrFWinterHet al. Oligoprogressive non-Small-Cell lung cancer under treatment with PD-(L)1 inhibitors. Cancers (Basel) (2020) 12:1046. doi: 10.3390/cancers12041046

  • CrossRef
  • Google Scholar

• 11

  SchoenfeldAJRizviHBandlamudiCSauterJLTravisWDRekhtmanNet al. Clinical and molecular correlates of PD-L1 expression in patients with lung adenocarcinomas. Ann Oncol (2020) 31:599–608. doi: 10.1016/j.annonc.2020.01.065

  • CrossRef
  • Google Scholar

• 12

  BudcziesJKazdalDAllgäuerMChristopoulosPRempelEPfarrNet al. Quantifying potential confounders of panel-based tumor mutational burden (TMB) measurement. Lung Cancer (2020) 142:114–9. doi: 10.1016/j.lungcan.2020.01.019

  • CrossRef
  • Google Scholar

• 13

  BudcziesJAllgäuerMLitchfieldKRempelEChristopoulosPKazdalDet al. Optimizing panel-based tumor mutational burden (TMB) measurement. Ann Oncol (2019) 30:1496–506. doi: 10.1093/annonc/mdz205

  • CrossRef
  • Google Scholar

• 14

  KazdalDEndrisVAllgäuerMKriegsmannMLeichsenringJVolckmarA-Let al. Spatial and temporal heterogeneity of panel-based tumor mutational burden in pulmonary adenocarcinoma: Separating biology from technical artifacts. J Thor Oncol (2019) 14:1935–47. doi: 10.1016/j.jtho.2019.07.006

  • CrossRef
  • Google Scholar

• 15

  BudcziesJKirchnerMKluckKKazdalDGladeJAllgäuerMet al. A gene expression signature associated with b cells predicts benefit from immune checkpoint blockade in lung adenocarcinoma. Oncoimmunology (2021) 10:1860586. doi: 10.1080/2162402X.2020.1860586

  • CrossRef
  • Google Scholar

• 16

  HelminkBAReddySMGaoJZhangSBasarRThakurRet al. B cells and tertiary lymphoid structures promote immunotherapy response. Nature (2020) 577:549–55. doi: 10.1038/s41586-019-1922-8

  • CrossRef
  • Google Scholar

• 17

  GaissmaierLChristopoulosP. Immune modulation in lung cancer: Current concepts and future strategies. Respiration (2020) 99:1–27. doi: 10.1159/000510385

  • CrossRef
  • Google Scholar

• 18

  Honrubia-PerisBGarde-NogueraJGarcía-SánchezJPiera-MolonsNLlombart-CussacAFernández-MurgaML. Soluble biomarkers with prognostic and predictive value in advanced non-small cell lung cancer treated with immunotherapy. Cancers (Basel) (2021) 13:4280. doi: 10.3390/cancers13174280

  • CrossRef
  • Google Scholar

• 19

  MountziosGSamantasESenghasKZervasEKrisamJSamitasKet al. Association of the advanced lung cancer inflammation index (ALI) with immune checkpoint inhibitor efficacy in patients with advanced non-small-cell lung cancer. ESMO Open (2021) 6:100254. doi: 10.1016/j.esmoop.2021.100254

  • CrossRef
  • Google Scholar

• 20

  WesselsSMuleyTChristopoulosPMeisterMHeinzmann-GrothIWarthAet al. Comprehensive serial biobanking in advanced NSCLC: feasibility, challenges and perspectives. Transl Lung Cancer Res (2020) 9:1000–14. doi: 10.21037/tlcr-20-137

  • CrossRef
  • Google Scholar

• 21

  VolckmarA-LLeichsenringJKirchnerMChristopoulosPNeumannOBudcziesJet al. Combined targeted DNA and RNA sequencing of advanced NSCLC in routine molecular diagnostics: Analysis of the first 3,000 Heidelberg cases. Int J Cancer (2019) 145:649–61. doi: 10.1002/ijc.32133

  • CrossRef
  • Google Scholar

• 22

  BartlettCHMardekianJCotterMHuangXZhangZParrinelloCMet al. Abstract P3-17-03: Concordance of real world progression free survival (PFS) on endocrine therapy as first line treatment for metastatic breast cancer using electronic health record with proper quality control versus conventional PFS from a phase 3 trial. Cancer Res (2018) 78:P3–17-03-P3-17-03. doi: 10.1158/1538-7445.SABCS17-P3-17-03

  • CrossRef
  • Google Scholar

• 23

  MaXNussbaumNCMageeKBourlaABTuckerMBellomoLet al. Comparison of real-world response rate (rwRR) to RECIST-based response rate in patients with advanced non-small cell lung cancer (aNSCLC). Ann Oncol (2019) 30:v651. doi: 10.1093/annonc/mdz260.103

  • CrossRef
  • Google Scholar

• 24

  U.S. Department of Health and Human Services. Common terminology criteria for adverse events (CTCAE) | protocol development | CTEP (Maryland, USA: National Institutes of Health. National Cancer Institute) (2022).

  • Google Scholar

• 25

  ZhangJ-MAnJ. Cytokines, inflammation, and pain. Int Anesthesiol Clin (2007) 45:27–37. doi: 10.1097/AIA.0b013e318034194e

  • CrossRef
  • Google Scholar

• 26

  AderkaDEnglemannHHornikVSkornickYLevoYWallachDet al. Increased serum levels of soluble receptors for tumor necrosis factor in cancer patients. Cancer Res (1991) 51:5602–7. Available at: https://aacrjournals.org/cancerres/article/51/20/5602/497130/Increased-Serum-Levels-of-Soluble-Receptors-for.

  • Google Scholar

• 27

  AderkaDWysenbeekAEngelmannHCopeAPBrennanFMoladYet al. Correlation between serum levels of soluble tumor necrosis factor receptor and disease activity in systemic lupus erythematosus. Arthritis Rheum (1993) 36:1111–20. doi: 10.1002/art.1780360812

  • CrossRef
  • Google Scholar

• 28

  CouperKNBlountDGRileyEM. IL-10: the master regulator of immunity to infection. J Immunol (2008) 180:5771–7. doi: 10.4049/jimmunol.180.9.5771

  • CrossRef
  • Google Scholar

• 29

  FrezzettiDGalloMMaielloMRD’AlessioAEspositoCChicchinelliNet al. VEGF as a potential target in lung cancer. Expert Opin Ther Targets (2017) 21:959–66. doi: 10.1080/14728222.2017.1371137

  • CrossRef
  • Google Scholar

• 30

  AliperAMFrieden-KorovkinaVPBuzdinARoumiantsevSAZhavoronkovA. A role for G-CSF and GM-CSF in nonmyeloid cancers. Cancer Med (2014) 3:737–46. doi: 10.1002/cam4.239

  • CrossRef
  • Google Scholar

• 31

  QianQZhanPYuLShiYChengJWeiSet al. Baseline levels and decrease in serum soluble intercellular adhesion molecule-1 during chemotherapy predict objective response and survival in patients who have advanced non-small-cell lung cancer. Clin Lung Cancer (2011) 12:131–7. doi: 10.1016/j.cllc.2011.03.009

  • CrossRef
  • Google Scholar

• 32

  CrinòLMetroG. Therapeutic options targeting angiogenesis in nonsmall cell lung cancer. Eur Respir Rev (2014) 23:79–91. doi: 10.1183/09059180.00008913

  • CrossRef
  • Google Scholar

• 33

  PerroneFBelluominiLMazzottaMBianconiMDi NoiaVMeacciFet al. Exploring the role of respiratory microbiome in lung cancer: A systematic review. Crit Rev Oncol Hematol (2021) 164:103404. doi: 10.1016/j.critrevonc.2021.103404

  • CrossRef
  • Google Scholar

• 34

  DengZLiZSunCXieHChenZLiuJet al. The association between inflammation, the microbiome and urethane-induced pulmonary adenocarcinoma. Oncol Lett (2018) 15:6352–60. doi: 10.3892/ol.2018.8167

  • CrossRef
  • Google Scholar

• 35

  ShengJXuZ. Three decades of research on angiogenin: a review and perspective. Acta Biochim Bbiophys Sin (2016) 48:399–410. doi: 10.1093/abbs/gmv131

  • CrossRef
  • Google Scholar

• 36

  HuangY-SChenJL-YChenH-MYehL-HShihJ-YYenR-Fet al. Assessing tumor angiogenesis using dynamic contrast-enhanced integrated magnetic resonance-positron emission tomography in patients with non-small-cell lung cancer. BMC Cancer (2021) 21:348. doi: 10.1186/s12885-021-08064-4

  • CrossRef
  • Google Scholar

• 37

  BratmanSVYangSYIafollaMALiuZHansenARBedardPLet al. Personalized circulating tumor DNA analysis as a predictive biomarker in solid tumor patients treated with pembrolizumab. Nat Cancer (2020) 1:873–81. doi: 10.1038/s43018-020-0096-5

  • CrossRef
  • Google Scholar

• 38

  DietzSChristopoulosPYuanZAngelesAKGuLVolckmarA-Let al. Longitudinal therapy monitoring of ALK-positive lung cancer by combined copy number and targeted mutation profiling of cell-free DNA. EBioMedicine (2020) 62:103103. doi: 10.1016/j.ebiom.2020.103103

  • CrossRef
  • Google Scholar

• 39

  ZulatoEAttiliIPavanANardoGDel BiancoPBoscolo BragadinAet al. Early assessment of KRAS mutation in cfDNA correlates with risk of progression and death in advanced non-small-cell lung cancer. Br J Cancer (2020) 123:81–91. doi: 10.1038/s41416-020-0833-7

  • CrossRef
  • Google Scholar

• 40

  AngelesAKChristopoulosPYuanZBauerSJankeFOgrodnikSJet al. Early identification of disease progression in ALK-rearranged lung cancer using circulating tumor DNA analysis. NPJ Prec Oncol (2021) 5:100. doi: 10.1038/s41698-021-00239-3

  • CrossRef
  • Google Scholar

• 41

  RussoAFranchinaTRicciardiGBattagliaASchifanoSSchifillitiMet al. Dynamic changes of neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and lactate dehydrogenase (LDH) during treatment with immune checkpoint inhibitors (ICIs) in non-small cell lung cancer (NSCLC). J Clin Oncol (2019) 37:2596. doi: 10.1200/JCO.2019.37.15_suppl.2596

  • CrossRef
  • Google Scholar

• 42

  IwaseSNakadaT-AHattoriNTakahashiWTakahashiNAizimuTet al. Interleukin-6 as a diagnostic marker for infection in critically ill patients: A systematic review and meta-analysis. Am J Emerg Med (2019) 37:260–5. doi: 10.1016/j.ajem.2018.05.040

  • CrossRef
  • Google Scholar

• 43

  FrieriM. Corticosteroid effects on cytokines and chemokines. Allergy Asthma Proc (1999) 20:147–59. doi: 10.2500/108854199778553082

  • CrossRef
  • Google Scholar

• 44

  SolanaRTarazonaRAielloAEAkbarANAppayVBeswickMet al. CMV and immunosenescence: from basics to clinics. Immun Ageing (2012) 9:23. doi: 10.1186/1742-4933-9-23

  • CrossRef
  • Google Scholar

• 45

  PaselloGFabricioASDel BiancoPFavarettoASalizzatoVPiccinLet al. Circulating cytokines as predictors of response to immune checkpoint inhibitors (ICIs) in patients (pts) with melanoma (Mel) and non–small cell lung cancer (NSCLC). J Clin Oncol (2022) 40:2549. doi: 10.1200/JCO.2022.40.16_suppl.2549

  • CrossRef
  • Google Scholar

• 46

  WangZZhanPLvYShenKWeiYLiuHet al. Prognostic role of pretreatment neutrophil-to-lymphocyte ratio in non-small cell lung cancer patients treated with systemic therapy: a meta-analysis. Transl Lung Cancer Res (2019) 8:214–26. doi: 10.21037/tlcr.2019.06.10

  • CrossRef
  • Google Scholar

• 47

  Da KangHParkC-KChungCOhI-JKimY-CParkDet al. Baseline serum interleukin-6 levels predict the response of patients with advanced non-small cell lung cancer to PD-1/PD-L1 inhibitors. Immune Netw (2020) 20:e27. doi: 10.4110/in.2020.20.e27

  • CrossRef
  • Google Scholar

• 48

  SchalperKACarletonMZhouMChenTFengYHuangS-Pet al. Elevated serum interleukin-8 is associated with enhanced intratumor neutrophils and reduced clinical benefit of immune-checkpoint inhibitors. Nat Med (2020) 26:688–92. doi: 10.1038/s41591-020-0856-x

  • CrossRef
  • Google Scholar

• 49

  WuLXieSWangLLiJHanLQinBet al. The ratio of IP10 to IL-8 in plasma reflects and predicts the response of patients with lung cancer to anti-PD-1 immunotherapy combined with chemotherapy. Front Immunol (2021) 12:665147. doi: 10.3389/fimmu.2021.665147

  • CrossRef
  • Google Scholar

• 50

  KotteasEABoulasPGkiozosITsagkouliSTsoukalasGSyrigosKN. The intercellular cell adhesion molecule-1 (icam-1) in lung cancer: implications for disease progression and prognosis. Anticancer Res (2014) 34:4665–72. Available at: https://ar.iiarjournals.org/content/34/9/4665.

  • Google Scholar

• 51

  ShibakiRMurakamiSShinnoYMatsumotoYYoshidaTGotoYet al. Predictive value of serum VEGF levels for elderly patients or for patients with poor performance status receiving anti-PD-1 antibody therapy for advanced non-small-cell lung cancer. Cancer Immunol Immunother (2020) 69:1229–36. doi: 10.1007/s00262-020-02539-2

  • CrossRef
  • Google Scholar

• 52

  ChristopoulosPPfeiferDBartholoméKFolloMTimmerJFischPet al. Definition and characterization of the systemic T-cell dysregulation in untreated indolent b-cell lymphoma and very early CLL. Blood (2011) 117:3836–46. doi: 10.1182/blood-2010-07-299321

  • CrossRef
  • Google Scholar

• 53

  MorraRdePPPietroluongoEOttavianoMTortoraMMucciBet al. Immunological signature of patients with thymic epithelial tumors. J Clin Oncol (2022) 40:8589. doi: 10.1200/JCO.2022.40.16_suppl.8589

  • CrossRef
  • Google Scholar

• 54

  WangHZhouFZhaoCChengLZhouCQiaoMet al. Interleukin-10 is a promising marker for immune-related adverse events in patients with non-small cell lung cancer receiving immunotherapy. Front Immunol (2022) 13:840313. doi: 10.3389/fimmu.2022.840313

  • CrossRef
  • Google Scholar

• 55

  LimSYLeeJHGideTNMenziesAMGuminskiACarlinoMSet al. Circulating cytokines predict immune-related toxicity in melanoma patients receiving anti-PD-1-Based immunotherapy. Clin Cancer Res (2019) 25:1557–63. doi: 10.1158/1078-0432.CCR-18-2795

  • CrossRef
  • Google Scholar

• 56

  StefaniaSColiaRCinziaRCorradoACantatoreFP. Off-label use of anti-IL-1 drugs in rheumatic diseases. Int J Immunopathol Pharmacol (2021) 35:20587384211006584. doi: 10.1177/20587384211006584

  • CrossRef
  • Google Scholar

• 57

  RidkerPMMacFadyenJGThurenTEverettBMLibbyPGlynnRJet al. Effect of interleukin-1β inhibition with canakinumab on incident lung cancer in patients with atherosclerosis: exploratory results from a randomised, double-blind, placebo-controlled trial. Lancet (2017) 390:1833–42. doi: 10.1016/S0140-6736(17)32247-X

  • CrossRef
  • Google Scholar

• 58

  LythgoeMPPrasadV. Repositioning canakinumab for non-small cell lung cancer-important lessons for drug repurposing in oncology. Br J Cancer (2022) 127:785–7. doi: 10.1038/s41416-022-01893-5

  • CrossRef
  • Google Scholar

• 59

  BarreraLMontes-ServínEBarreraARamírez-TiradoLASalinas-ParraFBañales-MéndezJLet al. Cytokine profile determined by data-mining analysis set into clusters of non-small-cell lung cancer patients according to prognosis. Ann Oncol (2015) 26:428–35. doi: 10.1093/annonc/mdu549

  • CrossRef
  • Google Scholar

• 60

  AngelesAKJankeFBauerSChristopoulosPRiedigerALSültmannH. Liquid biopsies beyond mutation calling: Genomic and epigenomic features of cell-free DNA in cancer. Cancers (Basel) (2021) 13:5615. doi: 10.3390/cancers13225615

  • CrossRef
  • Google Scholar

• 61

  ChristopoulosPSchneiderMABozorgmehrFKuonJEngel-RiedelWKollmeierJet al. Large Cell neuroendocrine lung carcinoma induces peripheral T-cell repertoire alterations with predictive and prognostic significance. Lung Cancer (2018) 119:48–55. doi: 10.1016/j.lungcan.2018.03.002

  • CrossRef
  • Google Scholar

• 62

  MathewMZadeMMezghaniNPatelRWangYMomen-HeraviF. Extracellular vesicles as biomarkers in cancer immunotherapy. Cancers (Basel) (2020) 12:2825. doi: 10.3390/cancers12102825

  • CrossRef
  • Google Scholar

• 63

  RajakumarTHorosRJehnJSchenzJMuleyTPeleaOet al. A blood-based miRNA signature with prognostic value for overall survival in advanced stage non-small cell lung cancer treated with immunotherapy. NPJ Prec Oncol (2022) 6:19. doi: 10.1038/s41698-022-00262-y

  • CrossRef
  • Google Scholar

• 64

  WuJHuangWYinCCheungYHAbramsDFallonJTet al. Blood-based gene expression profiling to reveal potential response biomarkers for immunotherapy in advanced lung cancer. J Clin Oncol (2020) 38:e15155. doi: 10.1200/JCO.2020.38.15_suppl.e15155

  • CrossRef
  • Google Scholar