A cytokine composition and its application

Abstract

This invention discloses a predictive model for assessing the risk of severe cytokine release syndrome (SCR) in patients with hematologic malignancies after CAR-T cell therapy. The model includes a nomogram and a prognostic risk scoring formula to assess the risk of SCR after CAR-T cell infusion. The biomarkers are the peripheral blood cytokine concentrations during 24-48 hours after CAR-T cell infusion, including five cytokines: CX3CL1, GZMB, IL4, IL6, and PDGFAA. The nomogram in this invention is the first predictive model established applicable to assessing the risk of SCR in patients with leukemia, lymphoma, and multiple myeloma undergoing CAR-T cell therapy. The model is simple, intuitive, and easy to implement, effectively helping clinicians assess the risk of SCR after CAR-T cell therapy, thus providing theoretical guidance for patient monitoring and treatment.

Description

Technical Field

[0001] This invention relates to the field of biomedical technology, specifically to a cytokine composition and its application. Background Technology

[0002] Chimeric antigen receptor T (CAR-T) cell therapy is an effective approach for refractory / relapsed hematologic malignancies that have failed standard therapies, achieving complete remission rates of 83%, 54%, and 33% in patients with B-cell acute lymphoblastic leukemia, B-cell non-Hodgkin lymphoma, and multiple myeloma, respectively. However, the systemic immune system activation induced by CAR-T cell therapy can also lead to life-threatening cytokine release syndrome.

[0003] Cytokine release syndrome (CRS) is the most common side effect of CAR-T cell therapy, observed in most patients receiving CAR-T cell therapies targeting CD19, BCMA, CD20, CD22, CD7, CD123, and other T-cell-related therapies. CRS manifests as constitutional symptoms such as fever, fatigue, and anorexia, but can rapidly progress to hypotension, respiratory failure, shock, and organ dysfunction. The incidence of high-grade CRS can be as high as 46%, depending on the specific product and patient. Even severe CRS is reversible with early and aggressive management. Therefore, early prediction and timely intervention for severe CRS are crucial for CAR-T cell therapy.

[0004] Currently, the establishment of several reported predictive models for cytokine release syndrome is based on small samples and limited to CD19 CAR-T cell products. They have not been validated in CAR-T cell therapy targeting other targets ([1]Teachey DT, Lacey SF, Shaw P A, et al. Identification of Predictive Biomarkers for Cytokine Release Syndrome after Chimeric Antigen Receptor T-cell Therapy for Acute Lymphoblastic Leukemia[J]. Cancer Discov, 2016, 6(6): 664-79. [2]Hay KA, Hanafi LA, Li D, et al. Kinetics and biomarkers of severe cytokine release syndrome after CD19 chimeric antigen receptor-modified T-cell therapy[J]. Blood, 2017, 130(21): 2295-2306.). There is an urgent clinical need for predictive models applicable to predicting the risk of severe cytokine release syndrome in patients with hematologic malignancies receiving CAR-T cell therapy. Summary of the Invention

[0005] To address the aforementioned deficiencies in the prior art, this invention provides a cytokine composition and its application for predicting the risk of cytokine release syndrome in patients with hematologic malignancies after receiving CAR-T cell therapy.

[0006] Since cytokine release syndrome (CRS) is characterized by significant systemic inflammation and elevated levels of inflammatory cytokines, peripheral blood cytokine levels are effective biomarkers for predicting the risk of severe CRS. We developed a nomogram based on five cytokines infused 24–48 hours after CAR-T cell therapy to predict the occurrence of severe CRS in patients with hematologic malignancies, and externally validated it in two independent cohorts from six hospitals. Compared to previous models applicable only to CD19 CAR-T cell therapy, this nomogram exhibits superior discrimination and is applicable to patients with hematologic malignancies treated with multiple CAR-T cell products (such as CD19, humanized CD19, CD19 / 22, BCMA, and humanized BCMA).

[0007] A cytokine composition for predicting the risk of cytokine release syndrome in patients with hematologic malignancies after CAR-T cell therapy, wherein the cytokine composition comprises CX3CL1, GZMB, IL4, IL6, and PDGFAA.

[0008] The present invention further provides the use of the cytokine composition as a detection marker in the preparation of a kit for predicting the risk of cytokine release syndrome in patients with hematologic malignancies after receiving CAR-T cell therapy.

[0009] This invention further provides a kit for predicting the risk of cytokine release syndrome in patients with hematologic malignancies after CAR-T cell therapy. The kit determines the risk of cytokine release syndrome in patients with hematologic malignancies after CAR-T cell therapy by jointly detecting the concentrations (pg / ml) of cytokines CX3CL1, GZMB, IL4, IL6, and PDGFAA in a sample. The kit includes reagents for detecting each cytokine separately. Preferably, the sample is peripheral blood from patients with hematologic malignancies after CAR-T cell therapy. More preferably, the sample is peripheral blood from patients with hematologic malignancies 24-48 hours after CAR-T cell infusion.

[0010] This invention also provides a model for predicting the risk of cytokine release syndrome in patients with hematologic malignancies after CAR-T cell therapy. This model includes the concentrations (in pg / ml) of five cytokines—CX3CL1, GZMB, IL4, IL6, and PDGFAA—in peripheral blood during 24-48 hours after CAR-T cell infusion. The concentrations of these five cytokines in the samples are measured, and the risk score is calculated using the following formula:

[0011] Risk score = 1.19 × ln(CX3CL1) + 0.91 × ln(GZMB) + 0.74 × ln(IL6) - 1.53 × ln(IL4) - 0.91 × ln(PDGFAA) - 6.77

[0012] The risk score uses 0 as the cutoff value. When a patient's risk score is below 0, the patient is considered low-risk, while when the risk score is greater than or equal to 0, the patient is considered high-risk.

[0013] Preferably, the hematologic malignancy type is at least one of the following: multiple myeloma, B-cell acute lymphoblastic leukemia, or B-cell non-Hodgkin lymphoma.

[0014] CAR-T cell therapy is a CAR-T cell therapy that targets CD19, humanized CD19, CD19 / 22, CD20, BCMA, or humanized BCMA.

[0015] This invention also provides a nomogram for predicting the risk of cytokine release syndrome in patients with hematologic malignancies after CAR-T cell therapy, characterized by including the concentrations of five cytokines—CX3CL1, GZMB, IL4, IL6, and PDGFAA—in peripheral blood during 24-48 hours after CAR-T cell infusion.

[0016] The nomogram includes a score scale in the first row, with scores ranging from 0 to 100; rows two through six are the natural logarithms of the concentrations of five cytokines between 24 and 48 hours after CAR-T cell infusion, with vertical lines drawn upwards to correspond to a specific score in the first row; the seventh row is the patient's total score, obtained by adding the scores of the five indicators in rows two through six to the corresponding scores in the first row; the total score is then drawn downwards to correspond to the eighth row, which represents the probability of the patient developing severe cytokine release syndrome after CAR-T cell therapy.

[0017] Preferably, the hematologic malignancy type is at least one of the following: multiple myeloma, B-cell acute lymphoblastic leukemia, or B-cell non-Hodgkin lymphoma.

[0018] CAR-T cell therapy is a CAR-T cell therapy that targets CD19, humanized CD19, CD19 / 22, CD20, BCMA, or humanized BCMA.

[0019] This invention establishes, for the first time, a model based on the concentrations (pg / ml) of five cytokines (CX3CL1, GZMB, IL4, IL6, and PDGFA) in peripheral blood during 24-48 hours after CAR-T cell infusion. This model can assess and predict the risk of severe cytokine release syndrome in patients with hematologic malignancies receiving different CAR-T cell therapies, providing a strong reference for clinical treatment decisions. Attached Figure Description

[0020] Figure 1 The flowchart for constructing the prediction model for this invention is shown.

[0021] Figure 2 The following plots show the results of factor analysis for selecting the logistic regression with LASSO regularization associated with severe cytokine release syndrome. In the plot, (A) shows the selection of the tuning parameter (k) of the LASSO model by 10-fold cross-validation with minimum criteria. The optimal value of the LASSO tuning parameter (k) is indicated by the vertical line. The plot (B) shows the LASSO coefficients of the 18 potential factors.

[0022] Figure 3 A nodal chart constructed using a risk scoring model.

[0023] Figure 4 To evaluate the column line Figure 1 The calibration curves show the consistency of the data. From top to bottom, the figures (A), (B), and (C) represent the three cohorts: multiple myeloma (training set), multiple myeloma (multicenter validation set from 6 hospitals), and lymphoma / leukemia, respectively.

[0024] Figure 5 The decision curve analysis results for evaluating the clinical net benefit of the nomogram are shown in the figure. From top to bottom, (A), (B) and (C) represent three cohorts: multiple myeloma (training set), multiple myeloma (multicenter validation set from 6 hospitals), and lymphoma / leukemia, respectively. Detailed Implementation

[0025] This invention discloses the application of peripheral blood cytokines, a model for predicting the risk of severe cytokine release syndrome in patients with hematologic malignancies receiving CAR-T cell therapy, and its construction method. Those skilled in the art can refer to the content of this document and appropriately modify the process parameters to achieve the desired results. It is particularly important to note that all similar substitutions and modifications are obvious to those skilled in the art and are considered to be included in this invention. The methods and applications of this invention have been described through preferred embodiments. Those skilled in the art can clearly modify or appropriately change and combine the methods and applications described herein without departing from the content, spirit, and scope of this invention to realize and apply the technology of this invention.

[0026] This invention is the first to establish a model based on the concentrations (pg / ml) of five cytokines (CX3CL1, GZMB, IL4, IL6, and PDGFA) in peripheral blood during 24-48 hours after CAR-T cell infusion. This model can assess and predict the risk of severe cytokine release syndrome in patients with hematologic malignancies receiving different CAR-T cell therapies. The model is presented in the form of a nomogram, providing a strong reference for clinical treatment decisions.

[0027] To more concisely and clearly demonstrate the technical solution, objectives, and advantages of the present invention, the technical solution of the present invention is described in detail below with reference to specific embodiments. Unless otherwise specified, the technical means used in the embodiments are conventional means well known to those skilled in the art.

[0028] Example 1

[0029] An embodiment of the method for constructing a model to predict the risk of severe cytokine release syndrome in patients with hematologic malignancies receiving different CAR-T cell therapies, as described in this invention, includes the following steps:

[0030] (1) Enrolled population:

[0031] Eighty-seven patients with refractory / relapsed multiple myeloma who received BCMA CAR-T cell therapy at the First Affiliated Hospital of Zhejiang University School of Medicine were enrolled as the training set (Chictr.org number, ChiCTR1800017404). Fifty-nine patients with refractory / relapsed multiple myeloma who received CAR-T cell therapy from five centers—Xuzhou Medical University, Shanghai Changzheng Hospital, the First Affiliated Hospital of China Medical University, the Second Affiliated Hospital of Zhejiang University School of Medicine, and the Institute of Hematology and Blood Diseases Hospital of the Chinese Academy of Medical Sciences—were enrolled as the external validation set 1. Sixty-eight patients with refractory / relapsed B-cell acute lymphoblastic leukemia or B-cell non-Hodgkin lymphoma who received CAR-T cell therapy at the First Affiliated Hospital of Zhejiang University School of Medicine were enrolled as the external validation set 2. Figure 1 ).

[0032] (2) Indicator establishment:

[0033] Peripheral blood samples were collected from patients 24-48 hours after CAR-T cell infusion for testing 45 cytokines. Simultaneously, patient clinical characteristics such as age, sex, baseline white blood cell count, hemoglobin, platelets, ferritin, C-reactive protein, and creatinine were collected for model construction. Based on Lee criteria, the patients screened in step (1) were divided into two groups based on cytokine release syndrome: severe cytokine release syndrome (> grade 2) and mild cytokine release syndrome (≤ grade 2).

[0034] (3) Model building:

[0035] First, univariate analysis was used to screen indicators affecting severe cytokine release syndrome in patients with hematologic malignancies receiving CAR-T cell therapy. Pre-infusion hemoglobin levels and 17 cytokines were screened and subjected to Lasso regression analysis, yielding seven cytokines: MIP1α, CX3CL1, Flt3L, GZMB, IL4, IL6, and PDGFAA. Figure 2 ).

[0036] Next, stepwise multivariate logistic regression was used according to the Akaike Information Criterion (AIC) to screen for significant factors influencing the occurrence of severe cytokine release syndrome in patients with hematologic malignancies receiving CAR-T cell therapy. These factors included the concentrations (in pg / ml) of five cytokines: CX3CL1 (C-X3-C motif chemokine ligand 1), GZMB (granzyme B), IL4 (interleukin 4), IL6 (interleukin 6), and PDGFAA (platelet-derived growth factor alpha polypeptide). The risk score for these five cytokines was calculated as follows: Risk Score = 1.19 × ln(CX3CL1) + 0.91 × ln(GZMB) + 0.74 × ln(IL6) - 1.53 × ln(IL4) - 0.91 × ln(PDGFAA) - 6.77. The risk score uses 0 as the optimal cutoff value. When a patient's risk score is below 0, the patient is considered low-risk, while when the risk score is greater than or equal to 0, the patient is considered high-risk.

[0037] Based on the formula obtained from the previous logistic regression analysis, a nomogram can be plotted by calling the "rms" package to visually reflect the risk of patients developing severe cytokine release syndrome. Figure 3 ).

[0038] The vertical scale represents the screened independent clinical risk factors, and the horizontal scale represents the corresponding scores. The nomogram includes a score scale in the first row, with scores ranging from 0 to 100; rows two through six are the natural logarithms of the concentrations (in pg / ml) of five cytokines between 24 and 48 hours after CAR-T cell infusion, each corresponding to a score in the first row; the seventh row is the patient's total score, obtained by summing the scores of the five indicators in rows two through six corresponding to those in the first row; and the eighth row is the probability of developing severe cytokine release syndrome after CAR-T cell therapy.

[0039] Example 2

[0040] Model evaluation.

[0041] The AUC value of the ROC curve is achieved through Bootstrap sampling (one of the most commonly used repeated sampling methods, which involves repeatedly drawing different new sample groups from the original sample group with replacement) and is used to evaluate the model's discriminative ability.

[0042] Table 1. Nodal plot of AUC in internal and external validation queues.

[0043]

[0044]

[0045] The calibration curve is also implemented using the Bootstrap self-sampling method. The basic idea of ​​the calibration curve is to plot the actual incidence rate on the ordinate and the predicted incidence rate on the x-axis. Then, based on the nomogram model, the survival rate of all patients at a specified time point is predicted. Patients are divided into several nodes according to their incidence rate from low to high. The average predicted incidence rate and calibration point for each node are calculated. Connecting all calibration points with a smooth curve yields the prediction curve. The standard curve reflects the actual survival situation. The closer the prediction curve is to the standard curve, the higher the prediction accuracy of the model. The risk scoring model shows good consistency on both the training set and the two independent validation sets. Figure 4 ).

[0046] Decision curve analysis can evaluate the clinical applicability of a model by calculating its net benefit (NB) at different thresholds. In clinical practice, the application of nomograms on both the training and two independent validation sets demonstrated clinical benefit. Figure 5 ).

Claims

1. Use of a reagent for detecting a cytokine combination in the preparation of a kit for predicting the risk of cytokine release syndrome in a hematological malignancy patient after receiving CAR-T cell therapy; The cytokine combination is: CX3CL1, GZMB, IL4, IL6 and PDGFAA; The type of hematological malignancy is at least one of the following: multiple myeloma, B-cell acute lymphoblastic leukemia, B-cell non-Hodgkin lymphoma, The CAR-T cell therapy is a CAR-T cell therapy targeting CD19, humanized CD19, CD19 / 22, CD20, BCMA or humanized BCMA.

2. Use according to claim 1, characterized in that, The sample is the peripheral blood of a hematological malignancy patient after receiving CAR-T cell therapy.

3. Use according to claim 2, characterized in that, The sample is the peripheral blood of a hematological malignancy patient between 24-48 hours after receiving CAR-T cell therapy and CAR-T cell reinfusion.

4. Use according to claim 1, characterized in that, The concentrations of the five cytokines CX3CL1, GZMB, IL4, IL6 and PDGFAA in the peripheral blood during 24-48 hours after CAR-T cell reinfusion are included, the concentrations of the five cytokines in the sample are detected respectively, and the risk score is calculated using the following formula: Risk score = 1.19 x ln(CX3CL1) + 0.91 x ln(GZMB) + 0.74 x ln(IL6) - 1.53 x ln(IL4) - 0.91 x ln(PDGFAA) - 6.77, Wherein, the risk score takes 0 as the cutoff value, when the risk score of the patient is lower than 0, it is a low-risk patient, and when the risk score is greater than or equal to 0, the patient is a high-risk patient.

Images