Introduction
Lung cancer is a major cause of mortality, accounting for approximately 1.6 million deaths annually worldwide [
1,
2]. Clinical trials have demonstrated that the administration of anti-PD-1 or PD-L1 antibodies to patients with advanced NSCLC has shown beneficial clinical outcomes, such as improved clinical response and survival rates. Non-small-cell lung cancer (NSCLC), which accounts for 80–85% of the cases, and small-cell lung cancer (SCLC) are the two major subtypes of lung cancer. In less than a decade, immune checkpoint inhibitors have become the standard treatment for NSCLC. Nevertheless, recent clinical trials involving anti-PD-1 therapy have revealed that only 15–25% of patients with NSCLC responded to immune checkpoint blockade therapy alone, regardless of their PD-L1 expression [
3‐
5]. Predictive biomarkers may help to identify patients with lung cancer who may benefit from PD-1/PD-L1 immunotherapy. Numerous studies have certified PD-L1 (programmed death-ligand 1) and TMB (tumor mutational burden) as the most reliable biomarkers [
6]. However, to date, PD-L1 expression is the only validated predictive factor for NSCLC [
7]. This has some limitations. For instance, some patients respond to PD-1/PD-L1 immunotherapy even in the absence of PD-L1 expression. Traditional therapies, such as chemotherapy, radiotherapy, or targeted therapy, can significantly affect the expression pattern of PD-L1. As a matter of fact in most clinical situations, most of the patients had received chemoimmunotherapy instead of mono-immunotherapy regardless of PD-L1 expression. The KEYNOTE-407 study showed that first-line treatment with pembrolizumab combined with chemotherapy can significantly prolong OS and PFS in metastatic squamous cell carcinoma and all populations with PD-L1 levels. The 4-year follow-up results of the KEYNOTE-189 study also showed that regardless of PD-L1 expression levels, the combination of pembrolizumab and chemotherapy can bring survival benefits in adenocarcinoma of NSCLC. Other predictive biomarkers including TMB and gene expression profiling are still under investigation; however, these tumor-associated markers are costly and invasive. Hence, there is an urgent need for peripheral blood biomarkers to help identify responding patients with NSCLC in clinical situations, as these biomarkers would be more cost-effective and clinically convenient.
Previous studies have proposed that a number of peripheral blood markers could be indicative of PD-1/PD-L1 inhibition in lung cancer, such as the neutrophil–lymphocyte ratio (NLR); platelet-lymphocyte ratio (PLR); neutrophil count [
8], absolute monocyte count (AMC); absolute eosinophil count (AEC); and serum biomarkers such as lactate dehydrogenase (LDH), plasma-albumin (ALB), and C-reactive protein (CRP)[
9]. PD-1 or PD-L1 expression on certain immune cell populations has been identified as biomarkers in response to immunotherapy in lung cancer. Moreover, the T-cell receptor (TCR) repertoire diversity of peripheral PD-1
+ T cells is thought to predict clinical outcomes after immunotherapy in patients [
10]. Studies have suggested that serum cytokine levels, including IL-5 and IFN-γ, could be effective indicators for predicting the clinical efficacy and survival rates in patients with cancer undergoing anti-PD-1 blockade treatment, including those with lung cancer [
11]. For chemoimmunotherapy settings, immunological and nutritional markers such as NLR and C-reactive protein-albumin ratio could also be useful for predicting the outcomes [
12]. Also, high sPD-L1 concentration is a negative predictor of chemoimmunotherapy efficacy in patients with NSCLC [
13]. In neoadjuvant settings, baseline NLR, PLR, monocyte-to-lymphocyte ratio (MLR), and systemic immune-inflammation index (SII) are associated with major pathological response (MPR) in NSCLC patients receiving neoadjuvant chemoimmunotherapy [
14].
Here, we retrospectively analyzed the efficacy of PD-1 antibody therapy in locally advanced or metastatic NSCLC patients, and preliminarily explored the correlation between peripheral blood biomarkers and clinical efficacy thus enabling us to identify the population that would benefit most from PD-1-based therapy through a simple blood draw.
Materials and methods
Patient inclusion and exclusion criteria
This single-center, retrospective study was conducted at Nanjing Drum Tower Hospital, analyzing medical record data of 101 patients with stage IIIA–IV NSCLC treated with PD-1 antibody immunotherapy as a first-line or subsequent-line treatment, either as monotherapy or in combination with chemotherapy. The inclusion criteria were as follows: (1) An Eastern Cooperative Oncology Group performance status (ECOG PS) of 0–2, and receipt of ≥ 2 cycles of immunotherapy; (2) Diagnosis of stage IIIA–IV unresectable NSCLC; (3) Willingness to undergo blood collection before and after immunotherapy for serum cytokine testing and T cell phenotype analysis. The exclusion criteria were active autoimmune disease, severe infectious diseases, or systemic immunosuppression.
Treatments and study assessments
Patients were administered albumin paclitaxel (260 mg/m2) or pemetrexed (500 mg/m2) in combination with carboplatin (AUC 5), lobaplatin (30 mg/m2), or cisplatin (75 mg/m2), and a PD-1 antibody (200 mg) every three weeks. For maintenance therapy, patients were treated with the PD-1 antibody, either with or without albumin paclitaxel or pemetrexed, according to pathology phenotype. After every two rounds of therapy, the Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1) was employed to evaluate clinical efficacy. The criteria were as follows: complete response (CR): complete regression of the target lesions; partial response (PR): a reduction of more than 30% of the total target lesions; progressed disease (PD): a greater than 20% increase of the total target lesions; stable disease (SD): a reduction of less than 30% or an increase of less than 20% of the total target lesions; disease control rate (DCR): CR + PR + SD.
Sample collection and flow cytometry
Blood samples were collected from 88 patients with NSCLC before and after two cycles of PD-1 antibody treatment concurrently with the first radiographic evaluation. Briefly, 10 mL of blood was obtained by venipuncture and collected into sterile tubes containing ethylenediaminetetraacetic acid. These samples were then transferred to the research laboratory within 2 h. Blood serum was collected and prepared for a mixed sample solution, as illustrated by the CBA assay kit for the detection of IL-2, IL-4, IL-5, IL-6, IL-10, IL-13, IFN-γ, and TNF-α (BD Biosciences, USA). Subsequently, peripheral blood mononuclear cells (PBMCs) were purified by Ficoll-Plaque (GE, Chicago, IL, USA) centrifugation and labeled for analysis with mouse monoclonal antibodies specific to CD3, CD4, CD8, CD279, CD45RO, CD62LC, D127, and CD25 (BioLegend, San Diego, CA, USA). The percentage of the relevant immune cell populations was quantified using FlowJo software (FlowJo LLC, Ashland, OR, USA).
Statistical analysis
Statistical analysis was conducted using GraphPad Prism 9 software (GraphPad Software Inc., La Jolla, CA, USA) and the R programming language. The associations between cytokine levels, T-cell phenotype, and clinical response were analyzed using the paired t-test or the Wilcoxon test. Survival curves were calculated using the Kaplan–Meier analysis and compared using the unadjusted log-rank test. A p-value < 0.05 was considered statistically significant.
Discussion
PD-L1 is currently a known prognostic indicator for the efficacy of immunotherapy in NSCLC. However, even in patients who are PD-L1 negative, a combination of chemotherapy and immunotherapy has been found to be effective [
15]. This approach has led to an increased response rate and a higher overall survival (OS) rate. However, some patients still exhibit primary resistance to immune checkpoint blockade therapy, and some acquire resistance during the immunotherapy process. The exact mechanisms underlying the resistance to PD-1 blockade therapy are not yet well understood; however, they likely involve multiple factors including an abnormal gut microbiome composition [
16], parallel immune inhibitory pathways such as Tim-3 or Lag-3 [
17], and a loss of antigen-presentation capacity. To address this concern, research has been conducted to identify peripheral blood biomarkers that can signal a response to PD-1 blockade in patients with NSCLC.
Through this study, we suggested that the reduction of lymphocyte absolute counts in patients with PD may also help us to identify patients who may not benefit from immunotherapy even before image evaluations. Our analysis also uncovered that baseline blood serum cytokine levels could predict the effectiveness of anti-PD-1 blockade therapy in patients with NSCLC. We found that elevated levels of Th1 cytokines such as IFN-γ and TNF-α, as well as Th2 cytokines such as IL-5, IL-4, and IL-10, already existing in patients with NSCLC, could reliably predict clinical responses to PD-1 blockade therapies. Furthermore, our results showed a positive correlation between higher baseline cytokine levels (> 10 ng/mL) and improved PFS in patients with advanced NSCLC undergoing immunotherapy. The role of IFNs as potent immunomodulators is important, guiding the actions on both innate and adaptive lymphocytes, augmenting natural killer cell cytotoxicity, and augmenting dendritic cell function—all of which are essential for the initiation of adaptive immune responses that inhibit tumor development [
18].
The recruitment and activation of neutrophils, macrophages, and lymphocytes at sites of damage, infection, and tumor development [
19] is a result of TNF-α and other proinflammatory factors. This is in agreement with the findings that higher IFN-
γ and TNF-
α cytokine levels at the time of diagnosis and three months after treatment initiation are linked to a better response to immunotherapy and a longer OS in NSCLC [
20]. Our results indicated that a preexisting Th1 cytokine response may help trigger a robust and long-lasting immune response after PD-1 antibody blockade. While tumor immunity is mainly governed by Th1-mediated cellular responses, we observed that preexisting higher levels of IL-5, IL-4, and IL-10 were also correlated with better clinical responses. Specifically, higher levels of IL-5 were associated with a significant prognostic effect. This may be attributed to the role of T helper-2 (Th2) lymphocytes and group 2 innate lymphoid cells (ILC2) in antibody production. These cells can boost the humoral immune response through the differentiation and growth of B cells, thereby promoting antibody response [
21].
The immune-inflammatory state of the patients may be indicative of their heightened sensitivity to immunotherapy, rather than an inadequate immune response. In patients with NSCLC who undergo anti-PD-1 treatment, the dynamic alterations of cytokine levels could not be used to predict clinical responses, even among those showing a stable or clinical response. We found that serum cytokines are unsuitable for disease monitoring, because the dynamic changes of cytokine levels may be affected by multiple factors including different treatments such as chemotherapy, the immune status of the patient, or treatment-related side effects. This aligns with previous findings that baseline serum cytokine levels could be used to predict the benefits of immunotherapy, but they are unsuitable for longitudinal disease monitoring [
22].
Accumulating evidence has suggested that CD8
+ T cells in the tumor microenvironment and systemic CD4
+ T-cell immunity have a significant role in maintaining antitumor responses. Previous research has shown that central memory CD4
+ T cells in peripheral blood can be used as predictors of PD-1 blockade therapy in patients with malignant melanoma [
23]. Previous studies have shown that CD70, the ligand for CD27, is essential for the successful priming of CD8
+ T cells and successful antitumor immunity [
24‐
30]. However, in our research, the majority of markers on CD4
+ T cells or CD8
+ T cells did not show any significant differences between responders and non-responders. In our previous study, we noticed a correlation between a greater proportion of CD45RO
+CD62L (central memory phenotype) to CD45RO
+CD62L (effecter memory phenotype) on CD4
+T cells and a more favorable clinical outcome of PD-1 blockade in patients with lung cancer, although the results were not statistically significant, possibly owing to the limited sample size. Even with an increased sample size in our recent study, these markers did not yield significant insights in predicting the PD-1 blockade outcomes in patients with NSCLC.
Previous research has suggested that certain biomarkers in the peripheral blood may be possible indicators of the effectiveness of PD-1 blockade immunotherapy in lung cancer. Of these, serum inflammation and nutritional markers are simple to assess in clinical settings and can be easily incorporated into clinical practice. Research suggests that pretreatment levels of AEC, AMC, ALB, NLR, and PLR were independent positive predictors of PD-1 inhibitors in patients with advanced NSCLC [
9]. In another study, patients with NLR values < 5, LDH levels < 240 U/L, or PNI ≥ 45 had significantly better outcomes. Multivariable analysis revealed that these parameters had an independent correlation with both enhanced PFS and OS [
8]. Some studies have indicated that dynamic changes of serum makers are associated with clinical outcomes. Patients whose NLR decreased six weeks after treatment tended to have a longer PFS, and similar results were found in repeated measurements [
31].
However, these studies are all retrospective and have a relatively small sample size. The predictive value of these blood serum markers on PFS or OS requires further validation by randomized studies with larger sample sizes and control groups. Moreover, there is an inconsistency in the cutoff values for these immune-inflammatory nutritional parameters across published studies, rendering the standards of evaluation challenging.
No significant treatment- or response-associated phenotypic differences were observed in bulk CD8 + T cells in several studies, which is in line with our results. Investigations into PD-1 expression on peripheral blood cells have demonstrated that it can enrich tumor-reactive cells in certain contexts [
32‐
34]. Additionally, research has indicated that the proliferation of PD-1
+ CD8
+ T cells after PD-1 targeted therapy may be linked to clinical outcomes [
35]. To further explore PD-L1 expression on peripheral blood cells, an exploratory study was conducted involving 32 patients with NSCLC undergoing PD-L1/PD-1 blockade therapies. A marked disparity in the proportion of PD-L1
+ CD11b
+ myeloid cells between objective responders and non-responders was observed; those with a PD-L1
+ CD11b
+ cell proportion above 30% initially demonstrated a response rate of 50% [
36]. Moreover, the expression of CD3
+CD8
+PD-1
+ cells in peripheral blood, when compared between DCR and PD patients, revealed a greater response to PD-1/PD-L1 blockade, indicating that the percentage of PD-1
+ cell populations among peripheral blood T cells could distinguish between objective responders and non-responders.
Given the minimal differences observed in the study, future research should focus on more specific cell populations, such as neoantigen-specific T cells. Studies showed that the diversity of tumor-antigen-related immune cells may also play a predictive role, in addition to the prevalence of PD-1
+ T cells. Studies have revealed that the diversity of peripheral PD-1
+CD8
+ TCR and the presence of neoantigen-specific CD8
+ T cells could predict the clinical benefits of anti-PD-1/PD-L1 therapy. Immunotherapy has caused a transformation in the TCR repertoire, with early changes in TCR clonality correlating with the immune response, thus influencing the clinical outcomes of anti-PD-1/PD-L1 treatment in NSCLC [
33,
37‐
40].
Our research has several restrictions. Firstly, it was a retrospective single-center study and the number of participants was limited, which underscores the need for paired pre- and post-treatment patient samples to better understand the dynamic changes. Secondly, we included all patients treated with PD-1 antibodies in combination with chemotherapy but did not establish a control group of patients who received only chemotherapy. Consequently, changes in cytokine levels or immune cell phenotypes might be influenced by the chemotherapeutic agents and are not only attributable to immunotherapy. Thirdly, the cutoff values of these parameters were not clearly defined; using different cutoff values could significantly influence the outcomes. In this study, 31.82% of patients were administered PD-1 inhibitors as their second or post-subsequent line of treatment. The baseline blood markers may have been affected by previous treatments including chemotherapy and other agents.
Despite these limitations, to the best of our knowledge, our study is the first to identify pretreatment Th1/Th2 cytokine levels in the peripheral blood of patients with NSCLC as potential indicators of response efficacy of and survival benefits from anti-PD-1/PD-L1 immunotherapy.
In conclusion, our study suggests that pretreatment serum cytokine levels could predict the clinical efficacy as well as PFS in patients with NSCLC undergoing anti-PD-1 blockade therapy. Patients showing reduced total lymphocyte counts after immunotherapy might experience poorer clinical outcomes. Moreover, a higher CD3+CD8+PD-1+T cell count in peripheral blood prior to treatment correlated with a more favorable objective clinical response. These biomarkers could help clinicians identify subpopulations that are more likely to benefit from anti-PD-1/PD-L1 immunotherapy therapy. The significance of these biomarkers warrants detailed investigation through extensive prospective studies in the future.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.