Immune biomarker combinations based on cd8 / foxp3 and pd-1 and pd-l1 and uses thereof

Abstract

The application discloses an immune biomarker combination based on CD8 / FOXP3 and PD-1 and PD-L1 and clinical application thereof. The application utilizes advanced multi-marker immunofluorescence (mIHC) detection and artificial intelligence analysis means to effectively combine and calculate four key immune markers, including immune checkpoint molecules PD-1 and PD-L1, anti-tumor T cell marker CD8 and immune inhibitory Treg marker FOXP3, and the expression of the four key immune markers in the special area of tumor tissues is used to guide clinical practice. The application has the advantages that a new effective combination detection and analysis method of key immune markers for predicting the prognosis of lung cancer patients is provided, and patients are stratified, which has potential value for guiding anti-PD-1 / PD-L1 treatment or Treg cell targeted treatment.

Description

Technical Field

[0001] This invention belongs to the field of biomedical science and technology related to tumor immunotherapy, and in particular relates to a combination of immune biomarkers based on CD8 / FOXP3, PD-1 and PD-L1 and their clinical applications. Background Technology

[0002] Lung cancer accounts for nearly one-fifth of all cancer deaths worldwide. Over the past decade, conventional treatments, such as resection, platinum-based chemotherapy, radiotherapy, and targeted therapy, have significantly improved the prognosis of non-small cell lung cancer (NSCLC). NSCLC may harbor oncogenic driver mutations, such as epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) mutations. However, only a subset of patients have these driver mutations, and acquired resistance is inevitable. Immune checkpoint-based immunotherapies, such as programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1), have recently provided a novel approach to NSCLC treatment, enabling longer overall survival (OS) in some patients, particularly those unresponsive to therapies targeting oncogenic driver mutations. Immune checkpoint inhibitors (ICIs), including pembrolizumab, nivolumab, atezolizumab, and durvalumab, have been approved by the FDA for the treatment of lung cancer. However, only approximately 20% of NSCLC patients benefit from ICI treatment.

[0003] The fundamental principle of immunotherapy is the regulation of tumor-immune interactions. Numerous reports have revealed the genetic, epigenetic, and transcriptomic characteristics of NSCLC patients, but our understanding of the NSCLC immune microenvironment remains incomplete. The classification of tumors as "hot" or "cold" based on the density and distribution of CD8+ T cells and FOXP3+ regulatory T (Treg) cells can predict clinical outcomes for various cancer patients, with "hot" tumors exhibiting potential sensitivity to immunotherapy. The practical applicability of CD8+ tumor-infiltrating lymphocytes (TILs) or FOXP3+ TILs as independent prognostic factors for NSCLC patients is controversial; therefore, investigating the balance between CD8+ TILs and FOXP3+ TILs in the tumor microenvironment is crucial. The prognostic role of the CD8 / FOXP3 ratio in the tumor microenvironment has been reported in various cancer types, including NSCLC; however, the association between a pre-existing CD8 / FOXP3 ratio in the tumor microenvironment and immunotherapy outcomes remains unclear. Given that PD-1 and PD-L1 have been used as therapeutic targets in clinical practice, combining these two biomarkers with CD8 / FOXP3 can better guide patient triage and immunotherapy.

[0004] Based on PD-L1 status and the number of tumor cells (TILs), tumors can be classified into four types: Type I adaptive immune resistance (PD-L1 positive, high TILs), Type II immune neglect (PD-L1 negative, low TILs), Type III intrinsic induction (PD-L1 positive, low TILs), and Type IV immune tolerance (PD-L1 negative, high TILs). However, detailed histological characteristics combining PD-1, PD-L1 with CD8 and FOXP3 have only been partially studied in NSCLC, despite these being predictors of immune-mediated tumor growth and regression. In recent years, the type, density, and spatial distribution characteristics of TILs in the local tumor microenvironment have provided a foundation for the development of immunotherapy for NSCLC patients, and several new techniques for assessing multiple biomarkers at the spatial and histological levels have been developed. These methods can better identify immunobiomarkers in the tumor immune microenvironment for prognostic stratification of NSCLC patients. In fact, the accuracy of multiplex immunohistochemistry (mIHC) / immunofluorescence (mIF) assays has been shown to be superior to PD-L1 expression and gene expression signatures in predicting the response of various tumor types to PD-1 checkpoint blockade. Summary of the Invention

[0005] Given the complexity and high heterogeneity of the tumor microenvironment and its impact on anti-tumor immunity and cancer immune evasion, the prognostic value of a single immune biomarker is limited. This invention integrates immune checkpoint molecule expression and tumor-associated immune cell distribution patterns for prognostic prediction in non-small cell lung cancer (NSCLC) patients. This invention analyzes tissue microarray (TMA) data from multiplex immunohistochemistry and measures the density of tumor-infiltrating CD8+, FOXP3+ immune cells, tumor cells (PanCK+), and programmed cell death 1 (PD-1)+ and programmed cell death ligand 1 (PD-L1)+ cells in both the peritumor and intratumor subregions. The results show that the density of CD8+, FOXP3+ immune cells infiltrating in the peritumor subregion is higher than that in the intratumor subregion. Furthermore, unsupervised hierarchical clustering analysis of these biomarkers suggests that a combination of a high CD8 / FOXP3 ratio, low PD-1 and PD-L1 immune checkpoint expression, and the absence of epidermal growth factor receptor (EGFR) mutations may be a favorable prognostic biomarker. On the other hand, based on cluster analysis, low CD8 / FOXP3 ratios and low expression of immune checkpoints (PD-1 and PD-L1) may be biomarkers for patients responding to treatment strategies targeting regulatory T cells (Tregs). Furthermore, an immune risk scoring model based on multivariate Cox regression was established, which can identify independent prognostic factors for NSCLC patients. The results of this invention suggest that, due to the complex interactions of different components and the heterogeneity of the immune microenvironment, the combined use of multiple factors based on a combination of biomarkers (CD8, FOXP3, PD-1, and PD-L1) may hold promise for predicting the prognosis and stratification of NSCLC patients.

[0006] This invention proposes a biomarker combination comprising four key immune biomarkers: the checkpoint molecules PD-1 and PD-L1, the anti-tumor T cell biomarker CD8, and the immunosuppressive Treg biomarker FOXP3. It also proposes the potential application of this biomarker combination in the prognosis of NSCLC. In a specific implementation, formalin-fixed paraffin-embedded (FFPE) tumor tissue microarrays (TMAs) from 98 NSCLC patients were evaluated using mIHC technology to explore the role of these biomarkers in patients. The data were used to assess disease prognosis, patient survival, and stratification. This invention found that compared to intratumoral subregions, the peritumoral subregion showed higher densities of infiltrating CD8+ and FOXP3+ immune cells, as well as higher levels of PD-1 and PD-L1 expression. The combination of these four key immune biomarkers in the peritumoral subregion can be used to predict patient prognosis. This invention employs unsupervised hierarchical cluster analysis to assess the expression of CD8 / FOXP3, PD-1, and PD-L1 immune checkpoints. Based on the expression of these biomarkers in specific regions of tumor tissue, it predicts patient prognosis, and patient stratification may guide anti-PD-1 / PD-L1 therapy or Treg cell targeting strategies. Furthermore, based on multivariate Cox regression, this invention establishes an immune risk scoring model, which can identify independent prognostic factors for NSCLC patients.

[0007] Immune risk scoring model: Risk score = 0.391 CD8+ (-0.374) FOXP3+(-0.396) PD1+0.272 PD-L1 +(-0.269) CD8 / FOXP3+ (-0.473) CD8 / PD1+0.181 CD8 / PD-L1

[0008] This invention proposes a combination of immunobiomarkers that combine the CD8 / FOXP3 ratio with PD-1 and PD-L1 levels.

[0009] In this invention, the CD8 / FOXP3 ratio combined with PD-1 and PD-L1 levels is used as a stratification method for NSCLC patients.

[0010] Furthermore, the immunobiomarker combination of the present invention also includes: deficiency of epidermal growth factor receptor (EGFR). That is, the present invention also proposes an immunobiomarker combination that combines the CD8 / FOXP3 ratio with PD-1 and PD-L1 levels and EGFR levels.

[0011] The present invention also proposes a risk scoring model based on CD8, FOXP3, PD-1, and PD-L1, and using CD8 / FOXP3, CD8 / PD-1, and CD8 / PD-L1 as factors of interest, which can be used as independent prognostic factors for NSCLC.

[0012] Cancer evolution is largely influenced by cell type, cell density, and immune cell location in tumor subregions; the distribution of immune checkpoints and tumor-associated immune cells affects patient prognosis. This invention analyzed tumor specimens obtained from a cohort of stage I-IV non-small cell lung cancer (NSCLC) patients, performing mIHC image-based analysis on PD-L1 expression in malignant cells and the density of tumor-associated lymphoid tissue (TILs) expressing CD8, PD-1, and FOXP3 in intratumoral and peritumoral compartments. The density of TILs expressing the immune markers of interest in the peritumoral region was significantly higher than that in the intratumoral compartments. Unsupervised hierarchical clustering analysis based on these markers (CD8 / FOXP3, PD-1, and PD-L1) identified three clusters with significantly different survival rates, and a small number of patients with high CD8 / FOXP3 and low PD-1 and PD-L1 expression, without EGFR mutations, showed better overall survival. Furthermore, a low CD8 / FOXP3 ratio and high immune checkpoint expression may indicate that NSCLC possesses a strong immune evasion capability. Based on the clustering results, patient stratification may provide guidance for anti-PD-1 / PD-L1 therapy: patients with low CD8 / FOXP3 ratios and low immune checkpoint expression may benefit from strategies targeting Treg cells.

[0013] Tumors can be categorized as "hot" or "cold" based on the level of immune cell infiltration, which can be used to predict clinical outcomes for various cancer patients. "Hot" indicates potential sensitivity to immunotherapy. Comprehensive analysis of the type, density, and location of immune cells in spatially distributed subregions may provide insights into the development of immunotherapy. Due to the heterogeneity in the distribution of expression markers of interest, analyzing the immunophenotype of individual subregions may be more valuable for prognosis. Data from this study show that the CD8 / FOXP3 ratio is higher in the peritumoral subregion than in the intratumoral subregion. These results suggest that the intratumoral region is more likely to be immunosuppressed than the peritumoral region, and this immunosuppressive state may influence different genetic characteristics of tumor cells. Recently, it was discovered that mutations in cytosolic isocitrate dehydrogenase (IDH1) can inhibit STAT1 signaling to induce CD8+ T cell accumulation, thereby promoting immune evasion in gliomas. This finding partially explains why the intratumoral immune microenvironment is more susceptible to immunosuppression than the peritumoral immune microenvironment.

[0014] The tumor immune microenvironment is spatially heterogeneous, with particularly pronounced differences between the tumor core and the infiltrative periphery. Hepatocellular carcinoma (HCC) studies report that T cells, B cells, and monocytes infiltrate the tumor core periphery and are associated with patient prognosis. The location of immune cells in CRC also has prognostic value, superior to and independent of other prognostic factors, and superior to traditional TNM staging. Consistent with this finding, some have suggested that subregion-specific enrichment of immune cells is a promising prognostic factor for NSCLC patients. Furthermore, it has been shown that the density of immune cells influences immunotherapy response. For example, PD-1+ cells are a potential biomarker for anti-PD-1 immunotherapy in certain cancers, such as head and neck cancer and HCC. Notably, this study proposes that the distribution of PD-1+ cells in NSCLC subregions exhibits high heterogeneity, which may lead to different responses to PD-1 blockade. Response to immunotherapy is typically dominant and coordinated by complex tumor-host-microenvironment interactions within the tumor periphery and intratumoral compartments.

[0015] Unsupervised hierarchical clustering analysis of the markers was used to define the correlations among the densities of CD8+ T cells, FOXP3+ Treg cells, PD-1+ cells, and PD-L1 cells. This study proposes that, compared to adjacent normal tissue, increased CD8+ T cell density is associated with increased FOXP3+, PD-1+, and PD-L1+ cell densities in tumor tissue. However, applying these markers alone did not show significant prognostic value in predicting patient survival. Higher CD8 / FOXP3 ratios were associated with more favorable prognoses for esophageal, rectal, and head and neck cancers.

[0016] This invention proposes the application of the aforementioned immunobiomarkers as biomarkers for non-small cell lung cancer (NSCLC). The applications include using a combination of CD8 / FOXP3, PD-1, and PD-L1 as targets to prepare diagnostic reagents for the occurrence and / or metastasis of NSCLC, or to prepare drugs for anti-tumor immunotherapy of NSCLC, or to prepare reagents for predicting survival time in NSCLC, or to prepare prognostic evaluation reagents for NSCLC.

[0017] In the application described in this invention, the combination of markers,

[0018] A high CD8 / FOXP3 ratio and low PD-1 and PD-L1 expression indicate a good prognosis for non-small cell lung cancer.

[0019] A low CD8 / FOXP3 ratio and high PD-1 and PD-L1 expression indicate that non-small cell lung cancer has a strong immune evasion ability and can benefit from anti-PD-1 or anti-PD-L1 treatment.

[0020] A low CD8 / FOXP3 ratio and low PD-1 and PD-L1 expression suggest that patients may benefit from targeted regulatory T-cell (Treg) therapy strategies.

[0021] In the application described in this invention, the density of FOXP3+ and CD8+ T cells in the peritumoral subregion is higher than their density in the intratumoral subregion.

[0022] When the density of CD8+ T cells increases, the density of FOXP3+ T cells, PD-1+ cells, and PD-L1+ cells in tumor tissue also increases.

[0023] In the application described in this invention, immune cells, proteins, and / or small molecules are prepared by targeting a combination of immune biomarkers CD8 / FOXP3, PD-1, and PD-L1.

[0024] In the application described in this invention, the sample used to determine the biomarker is obtained from the patient's tumor tissue or the subject's lung tissue.

[0025] In the application described in this invention, the levels of any one or more of the immune biomarkers CD8 / FOXP3, PD-1, and PD-L1 are determined at the protein and cellular levels.

[0026] In the application described in this invention, when the levels of any one or more of CD8 / FOXP3, PD-1, and PD-L1 are determined at the protein and cellular levels, the combination of CD8 / FOXP3, PD-1, and PD-L1 is determined by immunohistochemistry and / or tissue microarray, or by the tissue spatial distribution of the combination of CD8 / FOXP3, PD-1, and PD-L1; or by unsupervised hierarchical clustering analysis of the combination of CD8 / FOXP3, PD-1, and PD-L1.

[0027] In the application described in this invention, based on CD8, FOXP3, PD-1, and PD-L1, a risk scoring model is constructed using CD8, FOXP3, PD-1, PD-L1, CD8 / FOXP3, CD8 / PD-1, and CD8 / PD-L1 as factors of interest. An immune risk scoring model is established based on multivariate Cox regression, and the risk score is determined as an independent prognostic factor for non-small cell lung cancer patients.

[0028] The present invention also proposes an immunobiomarker combination detection reagent, which is a combination of CD8, FOXP3, PD-1 and PD-L1, which can be used in combination.

[0029] The immune biomarkers include a combination of CD8 / FOXP3, PD-1, and PD-L1.

[0030] The immune biomarker further includes: lack of epidermal growth factor receptor EGFR.

[0031] The present invention also proposes the application of the aforementioned immunobiomarker combination detection reagent in the preparation of diagnostic reagents for the occurrence and / or metastasis of non-small cell lung cancer, or in the preparation of reagents for predicting the survival time of non-small cell lung cancer, or in the preparation of prognostic evaluation reagents for non-small cell lung cancer.

[0032] This invention also proposes the use of the aforementioned immune biomarker combination detection reagent for stratifying patients with non-small cell lung cancer, guiding research on anti-PD-1 / PD-L1 therapy or targeting strategies for Treg cells.

[0033] The present invention also proposes an antibody that is a combination of CD8, FOXP3, PD-1 and PD-L1.

[0034] The present invention also proposes the application of the antibody in the preparation of diagnostic reagents for the occurrence and / or metastasis of non-small cell lung cancer, or in the preparation of drugs for the treatment of non-small cell lung cancer, or in the preparation of reagents for predicting the survival time of non-small cell lung cancer, or in the preparation of prognostic evaluation reagents for non-small cell lung cancer.

[0035] This invention also proposes the use of the antibody for stratifying patients with non-small cell lung cancer, guiding research on anti-PD-1 / PD-L1 therapy or targeting strategies for Treg cells.

[0036] This invention also proposes a detection reagent / kit comprising an antibody and / or a multilabel immunofluorescence (mIHC) reagent capable of specifically binding to the protein of the aforementioned immunobiomarker. The mIHC reagent is used to detect the expression of the aforementioned immunobiomarker in non-small cell lung cancer tumor tissue.

[0037] The immune biomarkers include a combination of CD8 / FOXP3, PD-1, and PD-L1.

[0038] The immune biomarker further includes: lack of epidermal growth factor receptor EGFR.

[0039] The present invention also proposes the application of the aforementioned detection reagent / kit in the preparation of diagnostic reagents for the occurrence and / or metastasis of non-small cell lung cancer, or in the preparation of drugs for the treatment of non-small cell lung cancer, or in the preparation of reagents for predicting the survival time of non-small cell lung cancer, or in the preparation of prognostic evaluation reagents for non-small cell lung cancer.

[0040] This invention also proposes the detection reagent / kit for stratifying patients with non-small cell lung cancer, guiding research on anti-PD-1 / PD-L1 therapy or targeting strategies for Treg cells.

[0041] This invention also proposes a method for screening candidate drugs for treating non-small cell lung cancer (NSCLC), alleviating or preventing the occurrence of NSCLC, and / or improving the prognosis of NSCLC. The method includes detecting the effect of the candidate drug on the levels of the immunobiomarkers or their downstream signals in a subject or a sample obtained from the subject. If, after using the candidate drug, the biological function of the immunobiomarker CD8 / FOXP3 increases and the biological function of the immunobiomarkers PD-1 and PD-L1 decreases, it indicates that the candidate drug has the effect of treating NSCLC, alleviating or preventing the occurrence of NSCLC, and / or improving the prognosis of NSCLC.

[0042] The increased biological function of the immune biomarkers includes increased expression levels of the immune biomarkers and / or enhanced downstream signaling pathways.

[0043] The decline in the biological function of the immune biomarkers includes a decrease or absence of expression of the immune biomarkers and / or a weakening of their downstream signaling pathways.

[0044] The immune biomarkers include a combination of CD8 / FOXP3, PD-1, and PD-L1.

[0045] The immune biomarker further includes: lack of epidermal growth factor receptor EGFR.

[0046] This invention also proposes a method for screening candidate drugs for treating non-small cell lung cancer (NSCLC), alleviating or preventing the occurrence of NSCLC, and / or improving the prognosis of NSCLC. The method includes detecting the levels of the aforementioned immune biomarkers in samples obtained from subjects, guiding anti-PD-1 / PD-L1 treatment or Treg cell targeting strategies through unsupervised hierarchical clustering analysis; or constructing a risk scoring model based on CD8, FOXP3, PD-1, PD-L1, CD8 / FOXP3, CD8 / PD-1, and CD8 / PD-L1 as factors of interest, establishing an immune risk scoring model based on multivariate Cox regression, and using the risk score as a method for prognostic assessment of NSCLC patients.

[0047] The immune risk scoring model: Risk score = 0.391 CD8+ (-0.374) FOXP3+(-0.396) PD1+0.272 PD-L1 +(-0.269) CD8 / FOXP3+ (-0.473) CD8 / PD1+0.181 CD8 / PD-L1

[0048] A high immune risk score indicates a poor prognosis for non-small cell lung cancer, while a low immune risk score indicates a better prognosis.

[0049] The immune biomarkers include a combination of CD8, FOXP3, PD-1, and PD-L1.

[0050] The present invention also proposes a method for diagnosing non-small cell lung cancer in subjects, or a method for preventing or treating non-small cell lung cancer in subjects in need, or a method for assessing the severity of non-small cell lung cancer in patients, or a method for predicting whether a patient will develop non-small cell lung cancer, or a method for predicting or prognostically assessing whether the patient is suitable for treatment targeting the immune biomarkers described above.

[0051] The method includes: determining the expression levels of immunobiomarkers, including immune checkpoint molecules PD-1 and PD-L1, anti-tumor T cell marker CD8, and immunosuppressive Treg marker FOXP3, in specific regions of tumor tissue from tumor tissue or lung tissue samples of the patient or subject, wherein the levels are used to prevent or treat non-small cell lung cancer in the subject, or to assess the severity of a patient with non-small cell lung cancer, or to predict whether a patient will develop non-small cell lung cancer, or to predict or assess whether the patient is suitable for treatment with targeted immunobiomarkers;

[0052] Alternatively, the levels of immune biomarkers in the lung tissue can be detected to prevent or treat non-small cell lung cancer in the subject, or to assess the severity of non-small cell lung cancer in patients, or to predict whether a patient will develop non-small cell lung cancer, or to predict or assess whether the patient is suitable for targeted immune biomarker therapy.

[0053] In the method described in this invention, the immune biomarkers include a combination of CD8 / FOXP3, PD-1, and PD-L1.

[0054] In the method of the present invention, the immune biomarker further includes: a deficiency of epidermal growth factor receptor EGFR.

[0055] In the method described in this invention, a high CD8 / FOXP3 ratio, low PD-1 and PD-L1 expression, and the absence of EGFR mutations indicate a good prognosis.

[0056] In the method described in this invention, a low CD8 / FOXP3 ratio and high PD-1 and PD-L1 expression indicate that NSCLC has a strong immune evasion ability.

[0057] In the method described in this invention, a low CD8 / FOXP3 ratio and low PD-1 and PD-L1 expression indicate benefit from a targeted regulatory T cell (Treg) therapy strategy.

[0058] In the method of the present invention, the diagnosis includes the following steps: i) determining the expression levels of immunobiomarkers in lung tissue obtained from the subject; ii) comparing the expression levels of immunobiomarkers in lung tissue in step i) with the average expression levels of immunobiomarkers; iii) when the expression levels of immunobiomarkers PD-1 and PD-L1 in lung tissue determined in step i) are more than twice the average expression levels of immunobiomarkers and CD8 / FOXP3 are more than twice the average expression levels of immunobiomarkers, the individual being measured is more likely to have non-small cell lung cancer biomarkers, the severity of non-small cell lung cancer biomarkers is more severe, and the patient is more suitable for a strategy of targeting immunobiomarkers to prevent or treat non-small cell lung cancer biomarkers.

[0059] The prevention or treatment includes the following steps: preventing or treating non-small cell lung cancer in a subject in need by targeting non-small cell lung cancer-specific biomarkers, including: administering a pharmaceutical composition to the subject, the pharmaceutical composition comprising a pharmaceutically acceptable carrier, an effective amount of a modulator of the non-small cell lung cancer-specific biomarker, and optionally an additional therapeutic agent, thereby preventing or treating non-small cell lung cancer; wherein the non-small cell lung cancer-specific biomarker is selected from a combination of CD8 / FOXP3, PD-1, and PD-L1.

[0060] In the method described in this invention, the modulator is selected from small molecule chemical agents, antisense oligonucleotides, small interfering RNA (siRNA), short hairpin RNA (shRNA), antibodies, and bioactive fragments or homologs of the antibodies.

[0061] This invention also proposes a method for detecting immune biomarkers in a subregion surrounding a tumor. The method includes optimizing an advanced mIHC assay to simultaneously evaluate five different markers to characterize the expression of CD8+ T cells, FOXP3+ Treg cells, PanCK+ tumor cells, and PD-1+ / PD-L1+ cells in a specific subregion. The method employs a pattern recognition-based algorithm that quantifies the expression levels of immune biomarkers in the subregion surrounding the tumor, enabling the detection of the described immune biomarkers in a specific subregion surrounding the tumor.

[0062] The immune biomarkers include a combination of CD8, FOXP3, PD-1, and PD-L1.

[0063] This invention is the first to propose combining the CD8 / FOXP3 ratio with the expression levels of PD-1 and PD-L1 as an immune biomarker. To date, there have been no reports of using this as a stratification method for NSCLC patients.

[0064] In the specific implementation plan, immune clustering based on immune cell density showed that patients could be divided into three subgroups with different immune cell distributions. Unsupervised hierarchical clustering analysis revealed that NSCLC patients with a higher CD8 / FOXP3 ratio and lower PD-1 and PD-L1 immune checkpoint expression (cluster 3) had a better prognosis, and this group of patients did not have EGFR mutations. PD-L1 overexpression is promoted by oncogenic and constitutive activation signals, including EGFR, Kirsten rat sarcoma virus oncogene homolog (KRAS), and protein kinase B (AKT), which is a mechanism of intrinsic tumor cell resistance. PD-L1 can be induced in cancer cells and immune cells (myelosuppressive cells, dendritic cells, macrophages, and lymphocytes) in the tumor microenvironment through inflammatory signals. This mechanism is an example of adaptive resistance. On the other hand, a low CD8 / FOXP3 ratio and high PD-1 / PD-L1 expression (cluster 1) may reflect a stronger immune evasion ability of NSCLC tumors. Patient stratification based on clustering strategies may guide anti-PD-1 / PD-L1 therapy or recommend Treg cell-targeted therapy for patients with low CD8 / FOXP3 ratios and low PD-1 / PD-L1 expression (cluster 2). Cox regression analysis was used to build a risk scoring model that considers immune variables in the peritumoral subregions. Patients with high-risk scores had higher lymph node stages and advanced disease, which may indicate that the prognosis of NSCLC patients depends on the complex interactions of peritumor components. Since the risk scoring model is derived from a mathematical algorithm, its biological significance needs to be defined. Further evaluation of more immune subsets, such as macrophages, B cells, and dendritic cells, is needed to gain deeper insights into the heterogeneous immune microenvironment in malignancies and to demonstrate clinical relevance.

[0065] Selecting patient biomarkers for treatment has always been challenging. PD-L1 expression, detected by immunohistochemistry, is one potential biomarker. To date, existing data show some conflicting results, but PD-L1 immunohistochemistry appears to have the potential for clinical application in selecting patients for anti-PD-1 or anti-PD-L1 therapy. The biology of the PD-1 / PD-L1 axis is complex, and clinical anti-PD-1 / PD-L1 therapy has demonstrated considerable heterogeneity; using PD-L1 levels alone to guide targeting strategies is not ideal. Based on the hierarchical clustering model of this invention, patient stratification based on clustering results may guide anti-PD-1 / PD-L1 therapy. Subgroup characteristics may also serve as prognostic predictors for NSCLC patients.

[0066] In this invention, cluster analysis based on markers of interest (ROIs) indicates that a high CD8 / FOXP3 ratio and low PD-1 and PD-L1 immune checkpoint expression, combined with the absence of EGFR mutations, may be favorable prognostic markers. In this study, clustering was performed based on CD8 / FOXP3 quantification and PD-1 and PD-L1 immune checkpoint expression, and patients were stratified based on these clusters to guide cancer immunotherapy (anti-PD-1 / PD-L1 therapy or Treg cell-targeted therapy). Furthermore, future research will focus on highly heterogeneous microenvironments to explore the spatial distribution of CD8, FOXP3, PD-1, and PD-L1 immune checkpoint expression, to better stratify patients and guide clinical immunotherapy for NSCLC patients. Attached Figure Description

[0067] Figure 1: Six-color mIHC characterization of tumor-associated immune cells and immune checkpoints in NSCLC tissues;

[0068] Figure (A) shows a multispectral image (MSI) of a TMA core from tissue of an NSCLC patient generated by digital scanning.

[0069] Figure (B) shows the cell phenotype used in this invention: the representative image was obtained using a Vectra Polaris microscope and shows multiple staining patterns in NSCLC tissue; scale bar: 200 μm.

[0070] Figure (C) shows the segmentation of the tumor into an intratumoral subregion, a stromal region, and a background region. The region where PanCK+ cells are located is considered the intratumoral subregion, and machine learning algorithms are used to distinguish each region. Scale bar: 200 μm.

[0071] Figure (D) shows a representative image of the subregion surrounding the tumor.

[0072] Figure 2: Comparison of cell densities of CD8+ T cells, FOXP3+ cells, PD-1+ cells and PD-L1+ cells in tumor and adjacent normal tissue, and the relationships between these markers;

[0073] Figure (A) is a heatmap of hierarchical clustering showing the density levels of CD8+, FOXP3+, PD-1+ and PD-L1+ cells, as well as a dendrogram of the unsupervised hierarchical clustering results of tumors and adjacent normal tissues.

[0074] Figure BD shows that the density of CD8+ cells, FOXP3+ Treg cells and PD-1+ cells in the tumor sample was significantly increased compared with the adjacent normal tissue (p < 0.01-0.0001).

[0075] Figure (E) shows that the expression of PD-L1 in tumor tissue tends to increase compared with adjacent normal tissue (p=0.067).

[0076] Figure (F) shows that the CD8 / FOXP3 ratio in the tumor tissue was significantly lower than that in the adjacent normal tissue (p < 0.0001).

[0077] Figure (GJ) shows that the densities of CD8+ T cells and FOXP3+ Treg cells were significantly higher in NSCLC tissues with high PD-1+ cell density (G, H) and high PD-L1 expression (I, J) (p < 0.05–0.0001).

[0078] Figure 3: NSCLC tissue samples show increased infiltration of T cell subsets in the peritumoral subregion;

[0079] Figure (A) is a hierarchical clustering heatmap showing the unsupervised hierarchical clustering results of the density of CD8+, FOXP3+, PD-1+ and PD-L1+ cells in the tumor subregion and the peritumoral subregion.

[0080] Figure (BC) shows the distribution of each immune cell subset in the peritumoral and intratumoral subregions.

[0081] The figure (DG) shows the differences in immune cell density according to the sampling strategy. The y-axis represents the density of each immune cell. The x-axis of each point is labeled with the sampling strategy peritumoral and intratumoral. A line connects two points of cell density in two subregions related to the same tumor. Paired t-tests were used to assess the differences, and p-values ​​are labeled in the figure.

[0082] Figure 4: Patient stratification based on CD8 / FOXP3, PD-1+, and PD-L1+ cell density has prognostic value;

[0083] Figure (AD) shows the Kaplan-Meier curves illustrating the prognostic impact of CD8, FOXP3, PD-1, and PD-L1 expression levels on overall survival (OS) in a sample of 97 NSCLC patients. High and low expression levels were distinguished by median. The log-rank test was used to determine significance.

[0084] Figure (E) shows the prognostic value of patient stratification based on the CD8 / FOXP3 ratio, PD-1+ cells, and PD-L1+ cells in the peritumor subregion.

[0085] Figure (F) shows the Kaplan-Meier curves illustrating the prognostic impact on overall survival (OS) for the three patient groups based on CD8 / FOXP3 ratio and checkpoint expression stratification (based on data from Figure 4E).

[0086] Figure 5: Kaplan-Meier analysis of the impact of CD8, CD8 / FOXP3, FOXP3, PD-1 and PD-L1 levels on overall survival (OS) in subgroups of patients;

[0087] Figure (A) shows the survival outcomes of 97 NSCLC patients with different CD8+ cell densities. These patients were from three subgroups generated based on the CD8 / FOXP3 ratio and PD-1 and PD-L1 levels. There was no significant difference in OS between patients with high or low CD8 expression (p > 0.05).

[0088] Figure (B) shows the survival outcomes of 97 NSCLC patients with different CD8 / FOXP3 ratios. Patients with higher CD8 / FOXP3 expression in cluster 3 had significantly increased survival rates (p < 0.05).

[0089] Figure (C) shows the survival outcomes of 97 NSCLC patients with different FOXP3 expressions. Patients with higher FOXP3 expression in cluster 3 had poorer survival rates (p < 0.05).

[0090] Among them, Figure (D) showed no difference in survival outcomes among patients with different PD-1 expression (p>0.05);

[0091] Figure (E) shows the survival outcomes of 97 NSCLC patients with different PD-L1 expression levels. Patients with high PD-L1 levels in cluster 3 had significantly increased survival (p < 0.05). The median was used to distinguish between high-density and low-density NSCLC. The log-rank test was used for statistical analysis.

[0092] Figure 6: Association between immunosuppressive state and EGFR or ALK mutations;

[0093] Figure (A) shows the distribution of EGFR-mutant patients in clusters 1-3; clusters 1 and 2 (immunosuppression) show a relationship with mutation status, but cluster 3 does not.

[0094] In Figure (B), ALK mutation patients are evenly distributed across all three clusters.

[0095] Figure 7 The establishment of risk scores and the assessment of their prognostic significance;

[0096] Figure (A) shows that the survival rate of the low-risk scoring group was significantly better than that of the high-risk scoring group.

[0097] Figure (B) shows the AUC of the prediction model.

[0098] Figure (C) shows a heatmap comparing immune cells and risk scores.

[0099] Figure (DG) shows the association between clinical information (D), age (E), tumor (F), lymph nodes (G), AJCC stage, and risk score. Detailed Implementation

[0100] The invention will be further described in detail below with reference to the specific embodiments and accompanying drawings. Except for the contents specifically mentioned below, the processes, conditions, and experimental methods for implementing the invention are all common knowledge and general knowledge in the art, and the invention does not have any particular limitations.

[0101] Example 1

[0102] Materials and methods

[0103] Patient cohort and tissue microarray preparation

[0104] The NSCLC tissues were obtained from Ruijin Hospital, affiliated with Shanghai Jiao Tong University School of Medicine. The use of human specimens was approved by the Ethics Committee of Ruijin Hospital, and informed consent was obtained from the patients participating in this study. These tissues were formalin-fixed and paraffin-embedded. H&E staining was performed on these sections, and the results for each section were reviewed by an independent surgical pathologist. Based on the H&E staining results, a TMA of 180 tissues was constructed, including 82 paired NSCLC tumor tissue samples and adjacent normal tissue samples (164 tissue samples), and an additional 16 NSCLC tumor tissue samples. One NSCLC tumor tissue sample and one adjacent normal tissue sample were excluded due to tissue incompleteness. This invention ultimately used 97 NSCLC tumor tissue samples and 81 adjacent normal tissue samples. The core diameter of each sample in the TMA was 1.5 mm.

[0105] Immunohistochemistry

[0106] Paraffin-embedded sections were dewaxed using xylene and gradient ethanol solutions. Endogenous peroxidase was neutralized with endogenous peroxidase blocking solution (Beyotime, Shanghai, China), and Opal antibody diluent / blocking agent (Akoya Biosciences, MA, USA) was used to block the binding of irrelevant antibodies. Primary antibodies included anti-CD8α (clone D8A8Y, dilution 1:500, CellSignaling, MA, USA), anti-FOXP3 (clone D2W8E™, dilution 1:100, Cell Signaling), anti-PD-1 (clone D4W2J, dilution 1:250, Cell Signaling), anti-PD-L1 (clone E1L3N®, dilution 1:350, Cell Signaling), and anti-PanCK (clone C11, dilution 1:50, Cell Signaling). Following the application of HRP anti-mouse / rabbit secondary antibody (Akoya Biosciences), to detect antibody staining, non-small cell lung cancer tissue microarrays were incubated with one of the following fluorophores (Opal Polaris 520, Opal Polaris 570, Opal Polaris 620, Opal Polaris 690, and Opal Polaris 780, diluted 1:100) according to the manufacturer's instructions. The tissue microarrays were then mounted in anti-quenching mounting media containing DAPI (Cell Signaling). A full-slide tissue scan at 20x magnification was performed using a Vectra Polaris system (AkoyaBiosciences) to capture stained images. To further analyze the density of immune cells in the peritumoral subregion, a 100 μm thick band was created outside the tumor-stromal junction. The peritumoral subregion was defined as the area 100 μm outside the tumor margin.

[0107] Data Analysis

[0108] Spectra were constructed using monochromatic slices, and image deconvolution was performed using inForm software (v2.4.8; Akoya Biosciences) according to the manufacturer's instructions. The settings for the training images were applied to batch analysis of all scanned images. Data tables exported from inForm 2.4.8 were further analyzed in RStudio (RStudio v3.6.1). (Following the manufacturer's instructions...) TM Akoya's instructions use the R packages "Phenoptr" and "PhenoptrReports" for analysis and export combined results of cell count, cell percentage, and cell density for further plotting.

[0109] Hierarchical cluster analysis

[0110] Based on the TMA results, hierarchical clustering was used to cluster NSCLC patients. Expression and ratios were transformed using log2, and hierarchical clustering was performed in R using Euclidean distance with the ward.D2 method.

[0111] Statistical analysis

[0112] Statistical analyses were performed using GraphPad Prism 8.0.2 software (GraphPad, Inc., San Diego, CA, USA). The t-test was used to determine statistical differences between two samples. Spearman correlation coefficients were used to assess the correlation between different marker expression levels. Multivariate Cox regression models were used to analyze independent prognostic factors and immune biomarkers. Kaplan-Meier regression was used to assess overall survival between subgroups, and the log-rank test was used to assess statistical significance. P < 0.05 was defined as statistically significant.

[0113] Experimental results

[0114] Using mIHC to explore NSCLC phenotypes

[0115] To characterize the resident and infiltrating immune cell landscape in NSCLC in situ, a mIHC workflow was established and optimized to simultaneously assess five different markers (CD8, FOXP3, PanCK, PD-1, PD-L1) to characterize CD8+ T cells, FOXP3+ Treg cells, PanCK+ tumor cells, and PD-1+ / PD-L1+ cells. After spectral splitting using inForm software, multicolor images were separated into individual channels, and the corresponding images were visualized (Figure 1A). This invention relies on 6-color mIHC staining of TMA samples to assess the density of CD8+ T cells, FOXP3+ Treg cells, and PD-1+ and PD-L1+ cells in tumor tissue and adjacent normal tissue (Figure 1B). inForm was then used to assess phenotypes in intratumoral and stromal regions, where intratumoral subregions were defined by a machine learning algorithm that identified PanCK-positive regions as intratumoral subregions (Figure 1C). Furthermore, an algorithm based on pattern recognition was designed to define a subregion surrounding the tumor (Figure 1D). Therefore, the levels of immune cells and checkpoints were assessed in different microscopic anatomical subregions.

[0116] Correlation between high-density tumor-associated immune cells and immunosuppressive phenotype in NSCLC

[0117] This invention initially compared the densities of CD8+ T cells, FOXP3+ Treg cells, and PD-1+ and PD-L1+ cells in tumor tissue and paired adjacent normal tissue. This invention is able to cluster samples into two groups based on their origin (tumor or adjacent normal tissue) (Figure 2A). This invention found an increased density of CD8+, FOXP3+, and PD-1+ cells in tumor tissue (p<0.01). Figure 2 (BD) There was no significant difference in PD-L1 cell levels between tumor tissue and adjacent normal tissue (Figure 2E). Next, this invention assessed the CD8+ to FOXP3+ cell ratio. Compared with tumor tissue, adjacent normal tissue showed a significantly higher CD8 / FOXP3 ratio (Figure 2F). This invention further analyzed the relationship between CD8+ T cells, FOXP3+ Treg cells, and PD-1+ and PD-L1+ cells in NSCLC. PD-1+ or PD-L1+ cells suppress T cell function and lead to local anti-tumor immunosuppression. Their high levels are generally closely associated with poor prognosis and poor patient survival in many cancers, including NSCLC. Therefore, this invention set out to determine whether there is any relationship between the levels of cells expressing PD-L1, PD-1, CD8, and FOXP3 (positive cell density per square millimeter or expression intensity per square millimeter). This invention first divided NSCLC samples into two groups based on median PD-1+ cell density. Then, the densities of CD8+ and FOXP3+ cells were compared between the two groups. It was found that the densities of CD8+ T cells and FOXP3+ Treg cells were significantly higher in the high PD-1 expression group (Figure 2G, H). When NSCLC samples were divided into two groups based on PD-L1 expression levels, this invention also observed significantly higher levels of CD8+ T cells and FOXP3+ Treg cells in the high PD-L1 expression group (Figure 2I, J). In summary, these results indicate that PD-1+ or PD-L1+ cells may influence tumor reactivity and immunosuppressive cell populations.

[0118] The peritumoral and intratumoral subregions of NSCLC samples exhibit distinct immune subsets and checkpoint features.

[0119] Due to the heterogeneity of tumor distribution, the prognostic value of immune variables needs to be analyzed separately in different subregions. This invention investigates whether the density levels of CD8+ T cells, FOXP3+ Tregs, PD-1+ cells, and PD-L1+ cells can refine NSCLC patient subgroups. Heatmaps and cluster analyses were performed to assess potential correlations between different NSCLC samples in the peritumoral and intratumoral subregions. This invention can distinguish between peritumoral and intratumoral subregions based on the levels of CD8+, FOXP3+, PD-1+, and PD-L1+ cells (Figure 3A). Direct comparisons between peritumoral and intratumoral subregions in patient samples are provided, and the proportions of cell subsets are analyzed (Figures 3B-C). By comparing immune subsets and checkpoints in the peritumoral and intratumoral subregions of the same patient, a significant increase in the proportions of CD8+, FOXP3+, and PD-1+ T cell populations was found in the peritumoral subregion (P < 0.01-0.0001). (Figures 3D-F). In contrast, the proportion of PD-L1+ cells did not differ significantly between the peritumoral and intratumoral subregions (Figure 3G). The high number of cytotoxic T cells in the peritumoral subregion may represent a response at the tumor periphery, prompting this invention to further investigate the peritumoral subregion.

[0120] Cluster analysis confirmed the prognostic value of considering both immune subsets and checkpoints in NSCLC.

[0121] This invention then evaluated the different prognostic significance of CD8+, FOXP3+, PD-1+, and PD-L1+ cell densities in the peritumoral subregion. Kaplan-Meier analysis of the entire patient cohort showed that considering CD8+ T cell, FOXP3+ Treg cell, PD-1+ cell, and PD-L1+ cell densities individually did not have significant prognostic value for patient survival. Figure 4 Various studies have proposed that tumors can be classified into different subtypes based on the density and spatial distribution of immune subsets in the peritumoral or intratumoral subregions (18). Here, unsupervised clustering based on immune cell density identified three subgroups with significantly different survival rates, with cluster 3 showing a significantly better overall survival (OS) than clusters 1 and 2 (P = 0.048) (Figure 4E-F). Cluster 3, with its high proportion of CD8+ cells and low proportion of PD-1 / PD-L1+ cells, was associated with a better prognosis. Furthermore, this invention may suggest that cluster 2 has a poorer prognosis, but this subgroup may be able to improve CD8+ cell function by targeting Treg cells. Cluster 1 may require anti-PD-1 / PD-L1 therapy.

[0122] High CD8 / FOXP3 and high PD-L1 were associated with better survival outcomes in a subgroup of patients.

[0123] This invention further evaluated three patient subgroups (i.e., Group 1: intermediate CD8 / FOXP3 ratio, high PD-1 and PD-L1; Group 2: low CD8 / FOXP3 ratio, low PD-1 and PD-L1; Group 3: high CD8 / FOXP3 ratio, low PD-1 and PD-L1) (Figures 5A-E). High levels of tumor-infiltrating CD8+ T cells are considered a good predictor of survival in many human cancers, including NSCLC. In this study, no significant correlation was found between tumor-infiltrating CD8+ T lymphocytes and patient survival in any patient or subgroup (Figures 4A, 5A). When assessing the prognostic value of the CD8 / FOXP3 ratio, a high CD8 / FOXP3 ratio was found to predict higher survival in cluster 3 patients, but not in other subgroups (Figure 5B). The results indicate that CD8 levels have prognostic value relative to FOXP3 levels in patient subgroups. Furthermore, patients with cluster 2 high FOXP3 expression (low CD8 / FOXP3 ratio) have poorer survival rates. Figure 5 (C) This may indicate that targeting FOXP3+ Treg cells can improve CD8+ cell function and benefit the prognosis of this subgroup of patients. PD-1 expression did not show prognostic value in any of the three subgroups (Figure 5D). Surprisingly, PD-L1 level showed prognostic value in cluster 3 but not in cluster 1, and high PD-L1 expression was associated with better survival outcomes (Figure 5E). These results may reflect that a high CD8 / FOXP3 ratio and low PD-1 and PD-L1 expression can balance the effects of tumor reactivity and the immunosuppressive microenvironment, thereby achieving better survival outcomes.

[0124] Immunosuppression is regulated by PD-1 / PD-L1, which is associated with EGFR status.

[0125] EGFR driver mutation-positive NSCLC tissues overexpress PD-L1. In lung adenocarcinoma, tumor PD-L1 expression is positively correlated with EGFR mutation status and poor prognosis. This invention further evaluated whether immunosuppression is regulated by PD-1 / PD-L1 expression and EGFR or ALK mutation status. Patients with immunosuppression in clusters 1 (18.92%) and 2 (14.63%) had more EGFR mutations than patients in cluster 3 (0%) without EGFR mutations. Figure 6A), but ALK mutations had no effect on any of the three: cluster 1 (16.22%), cluster 2 (17.07%), and cluster 3 (10.53%) (Figure 6B). This data from the present invention indicates that EGFR mutation status is associated with high TIL density and high PD-1 / PD-L1 expression (cluster 1) and low TIL density and low PD-1 / PD-L1 expression (cluster 2). This result echoes previous reports that NSCLC patients with EGFR mutations have high PD-L1 expression.

[0126] Establishment of an immune risk scoring model for prognosis based on multivariate Cox regression

[0127] Although immune clustering can predict postoperative survival in NSCLC, it does not provide a linear measure of risk. Therefore, a risk score model was constructed using factors of interest: risk score = 0.391 CD8+ (-0.374) FOXP3+(-0.396) PD1+0.272 PD-L1 +(-0.269) CD8 / FOXP3+ (-0.473) CD8 / PD1+0.181 CD8 / PD-L1. Notably, the survival rate of patients with high-risk scores was significantly lower than that of patients with low-risk scores (P=0.012); Figure 7 A). The immune markers included in this scoring model are located in the peritumoral subregion; therefore, the score reflects the prognosis of NSCLC based on the complex interactions of these components within the peritumoral subregion. Furthermore, to assess the sensitivity and specificity of the risk score in determining the prognosis of NSCLC patients, a time-dependent ROC analysis was performed (Figure 7B). ROC curve analysis showed that the AUC values ​​for 1-year, 3-year, and 5-year overall survival were 0.784, 0.698, and 0.722, respectively, with models having higher AUC values ​​performing better than those with lower AUC values. Overall, the risk scoring model based on the levels of immune markers in the peritumoral subregion effectively predicts overall survival in NSCLC patients. Figure 7C shows a heatmap of the correlation between immune cells and the risk score. This invention further analyzed the association between clinicopathological features (age, tumor stage, lymph node stage, and AJCC stage) and the risk score. Figure 7 DG showed that patients with higher lymph node stage (P=0.03) and AJCC stage (P=0.02) had significantly higher risk scores than patients with lower lymph node stage and AJCC stage, indicating that the risk scoring model has substantial prognostic value.

[0128] The scope of protection of this invention is not limited to the above embodiments. Any variations and advantages that can be conceived by those skilled in the art without departing from the spirit and scope of the inventive concept are included in this invention and are protected by the appended claims.

Claims

1. The use of an immune marker combination in constructing an immune risk score model for the prognosis of non-small cell lung cancer patients, characterized in that, The immune marker combination is anti-tumor T cell marker CD8, immune suppressive Treg marker FOXP3, immune checkpoint molecule PD1 and PD-L1 in the sub-region around the tumor, and the immune risk score model is: risk score = 0.391*(CD8 expression level) + (-0.374)*(FOXP3 expression level) + (-0.396)*(PD1 expression level) + 0.272*(PD-L1 expression level) + (-0.269)*(CD8 expression level) / (FOXP3 expression level) + (-0.473)*(CD8 expression level) / (PD1 expression level) + 0.181*(CD8 expression level) / (PD-L1 expression level).

Images