Liquid biopsy analysis of cellular states for predicting immunotherapy toxicity

Abstract

Disclosed herein are methods for predicting the likelihood of a melanoma patient developing a severe immune-related adverse effect (irAE) associated with administration of immunotherapy based on the abundance of activated CD4 memory T cells and / or the diversity of T cell receptors (TCRs) in a peripheral blood sample obtained from the patient.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to U.S. Provisional Application Serial No. 63 / 299,377, filed on January 13, 2022, which is incorporated herein by reference in its entirety. STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0003] This invention was made with government support under Grants CA187192 and CA238711 from the National Institutes of Health. The U.S. Government has certain rights in this invention.

[0004] Materials Incorporated by Reference

[0005] not applicable. Technical Field

[0006] The present invention generally relates to methods for predicting immunotherapy toxicity in a patient. Background Art

[0007] Although ICIs have revolutionized cancer treatment, severe immune-related toxicities currently occur in approximately 10%-60% of ICI-treated patients with melanoma, with toxicity rates closely related to the specific therapy administered. ICI-induced toxicities, also known as irAEs, affect a range of organ systems, including the lungs, liver, heart, skin, pituitary, and gastrointestinal tract, and can be associated with severe morbidity requiring urgent medical intervention. This morbidity can lead to the suspension of anticancer therapy and, in the most severe cases, death. The biological drivers of irAEs are not well characterized, and there is no method in standard clinical practice to identify which patients are at highest risk for developing irAEs.

[0008] Therefore, several groups have investigated potential biomarkers of ICI-induced toxicity based on blood or tumor analysis. However, these studies have typically focused on early prediction or single organ systems in treatment, with only general performance in predicting irAEs in the pretreatment setting independent of the affected organ system. Recently, a candidate irAE biomarker for pneumonia alone using tumor immunohistochemistry was reported; however, this biomarker was indirectly identified from the Cancer Genome Atlas, which lacks toxicity annotations and was evaluated in a case-control setting that did not include low-grade irAEs. Another group identified a single nucleotide polymorphism within the gene encoding microRNA-146a, which was associated with the development of severe irAEs. Despite this, other groups have identified ICI response biomarkers without examining irAEs.

[0009] Given the considerable heterogeneity of ICI-induced irAEs, including variations in their timing, severity, and location, identifying the factors that cause them remains challenging. Pre-existing autoantibodies, autoreactive tissue-resident T cells, and T cells specific for viral antigens derived from chronic viral infection have all been associated with irAEs. Alterations in the gut microbiome that lead to increased colonic interleukin-1β expression have also recently been reported in ICI-induced colitis. Given these observations, several groups have investigated the similarities between irAEs and autoimmune diseases. Indeed, case reports have shown that ICIs can cause overt autoimmunity, suggesting that irAEs may represent subclinical autoimmunity in subgroups of patients. However, it is unclear whether common immune states precede the different manifestations of ICI-induced toxicity. Summary of the invention

[0010] In various aspects of the present invention, there is provided a method and composition for predicting the possibility of serious immune-related adverse reactions (irAE) in patients receiving immunotherapy based on biomarkers derived from peripheral blood samples obtained from patients before receiving immunotherapy. In one aspect, the disclosed method includes obtaining peripheral blood samples from subjects before receiving immunotherapy treatment, and quantifying the abundance of activated CD4 memory T cells and the diversity of T cell receptors (TCR) in peripheral blood samples. The preferred method also includes that if the abundance of activated CD4 memory T cells and the combination of the diversity of T cell receptors (TCR) exceed a threshold value (sometimes referred to as a model index in this article), the patient is classified as a patient who may have severe irAR. The model index can be used to determine the threshold value, and the model index identifies the level and TCR diversity of activated CD4 memory T cells, and provides a series of values ​​representing an upper limit, which indicates that the patient is susceptible to irAR when it exceeds the upper limit. In one aspect, the value of the model index (a combination of CD4 memory T cells and TCR diversity values) exceeding a predetermined threshold value predicts more serious irAR. In another aspect, the method also includes determining the threshold value by reference to known clinical standards. In another aspect, the disclosed method comprises determining the abundance of activated CD4 memory T cells and the diversity of T cell receptors (TCRs) using at least one of: bulk RNA sequencing (CIBERSORTx and MiXCR), time-of-flight mass cytometry (CyTOF), TCR-β analysis, droplet-based scRNA-seq and scTCR-seq, and targeted RNA-seq using the RNAome targeting activated CD4 memory T cells.

[0011] In other aspects of the present invention, methods for predicting the likelihood of severe immune-related adverse reactions (irAEs) in patients receiving immunotherapy are disclosed. In one aspect, the method includes obtaining a first peripheral blood sample from a subject before receiving immunotherapy treatment, and obtaining a second peripheral blood sample after administering immunotherapy. The methods disclosed in these other aspects include quantifying a first TCR diversity level from a first peripheral blood sample and a second TCR diversity level from a second peripheral blood sample. The disclosed method also includes obtaining the degree of TCR amplification by subtracting the first TCR sub-diversity level from the second TCR diversity level. The disclosed method also includes classifying the patient as likely to have severe irAR if the degree of TCR amplification exceeds a threshold. In one aspect, these methods include predicting the onset time of severe irAR based on the degree of TCR amplification, wherein a higher degree of TCR amplification predicts an earlier onset of severe irAR. In one aspect, these methods include determining the level of T cell receptor (TCR) diversity using at least one of the following: bulk RNA sequencing (CIBERSORTx and MiXCR), time-of-flight mass cytometry (CyTOF), TCR-β analysis, droplet-based scRNA-seq and scTCR-seq, and targeted RNA-seq using the RNAome targeting activated CD4 memory T cells.

[0012] Other objects and features will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Those skilled in the art will appreciate that the drawings described below are for illustration purposes only and are not intended to limit the scope of the present teachings in any way.

[0014] Figure 1 is a schematic diagram of the study protocol described in the present invention, including an overview of the patients included in this study, a summary of the irAE status of the patients, exclusion criteria, and downstream analyses performed. Of the total 78 eligible patients, 71 were evaluable for irAE analysis after applying the exclusion criteria.

[0015] Figure 2A is a set of color-coded charts representing Figure 1 Characteristics of the single-cell discovery cohort, including the highest irAE grade experienced after initiation of immunotherapy and durable clinical response status.

[0016] Figure 2B UMAP plot of viSNE mapping of peripheral blood cells analyzed by CyTOF. t-SNE stands for t-distributed stochastic neighbor embedding.

[0017] Figure 2CFigure 2 is a heat map showing the relative abundance of 20 cell states identified by CyTOF in 18 patients grouped by future irAE status (left), and a graph showing the association of cell state abundance with the occurrence of severe irAEs (right). Statistical significance was determined by a two-sided unpaired Wilcoxon rank-sum test and expressed as directional -log10 P values. For associations without severe irAEs, -log10 P values ​​were multiplied by -1. Q values ​​were determined by the Benjamini-Hochberg method.

[0018] Figure 2D Figure 2 is a graph of the frequency of CD4 TEM cells (CyTOF) in peripheral blood before treatment in patients stratified by future irAE status (no severe irAE, n = 10 patients; with severe irAE, n = 8 patients). The box center line, box borders, and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively. Statistical significance was determined by a two-sided unpaired Wilcoxon rank sum test.

[0019] Figure 3A By CyTOF ( Figure 2A ) UMAP of peripheral blood cells from scRNA-seq analysis of 13 patients analyzed in total, colored by cell type, patient, and status (n=32). T / NKT indicates NK-like T cells.

[0020] Figure 3B is a UMAP of cell state abundance (scRNA-seq) versus future irAE status and CD4 TEM cell frequency (CyTOF). The former was quantified by a two-sided unpaired Wilcoxon rank sum test and expressed as a -log10 P value. For association with no severe irAEs, the -log10 P value was multiplied by -1. CD4 T cell states 5 and 3 are collectively expressed as CD4 T5+3.

[0021] Figure 3C is the DEG between CD4 T cell states 5 and 3 and other CD4 T cell states (P adj < 0.05). Within each state, the columns represent the mean expression of a single patient converted to a z-score.

[0022] Figure 3DIt is a set of 2 figures, (left) the frequency of activated and resting candidate subgroups of CD4 T 5+3 cell status in 13 patients stratified by non-severe (n=7) and severe (n=6) irAE status. Counts per million (CPM) of activation markers>0 indicates expression. Significance was determined by a bilateral unpaired Wilcoxon rank sum test, and (right) a receiver operating characteristic curve plot showing the performance of the CD4T 5+3 subgroup (from the left figure) (for predicting the development of severe irAEs). NS means not significant.

[0023] Figure 3E is a chart showing each T cell status, total T cells, CD8 T cells, CD4 T cells and activated vs. resting CD4 T 5+3 cells (eg. Figure 3D Figure 2. Plot of pre-treatment TCR clonotype diversity within irAEs defined by irAE status. TCR diversity was calculated for all patients (n=9) with at least 100 TCR clones. Status was ranked by AUC (between TCR diversity and severe irAE status).

[0024] Figure 3F Is a color-coded plot of the average expression of key lineage and activation genes in CD4 T cell states. States within the box are consistent with TEM and TEM-like phenotypes. The box center line, box borders, and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively.

[0025] Figure 4A The bulk cohort (n = 26 patients) and bulk cohort 2 (n = 27 patients) are shown ( Figure 1 Figure 3 Plot of the association between pre-treatment peripheral blood leukocyte composition (CIBERSORTx) and the occurrence of severe irAEs in patients with irAEs. Significance was determined by a two-sided unpaired Wilcoxon rank sum test and expressed as a -log10 P value. For associations without severe irAEs, the -log10 P value was multiplied by -1.

[0026] Figure 4B Figure 2 shows TCR clonotype diversity (Shannon entropy) in two bulk cohorts (n = 53 patients), with patients stratified by future irAE status (no severe irAE, n = 36; severe irAE, n = 17). The box center line, box borders, and whiskers indicate the median, first and third quartiles, and minimum and maximum values, respectively. Significance was determined by a two-sided unpaired Wilcoxon rank sum test.

[0027] Figure 4CFigure 2 is a graph showing the development of a composite model for predicting severe irAEs that integrates activated CD4 TM cell abundance and TCR clonotype diversity from pre-treatment peripheral blood transcriptome, with model scores trained for bulk cohort 1 and displayed in both cohorts. The cut-points for high / low scores were optimized for bulk cohort 1 using Youden’s J statistic.

[0028] Figure 4D It is a set of 2 figures, (left) ROC plots showing the composite model performance in bulk cohort 2 (hold-out validation) (whether applied to all patients (two therapies, n=27), combination therapy patients (n=11) or PD-1 monotherapy patients (n=16)), and (right) ROC plots showing the composite model performance in bulk cohorts 1 and 2 (whether trained on PD-1 patients (n=29) and tested on combination therapy patients (n=24) or vice versa). The AUC of each ROC curve is shown.

[0029] Figure 4E It is a graph of the composite model scores of all batch cohort patients (n=53) after model training with LOOCV for severe irAE occurrence (Figure 13), grouped by the highest irAE level of each patient. The center line of the box, the box boundary, and the whiskers represent the median, the first quartile, and the third quartile, and the minimum and maximum values ​​within the quartile range of the 1.5x box limit, respectively. Statistical significance was determined by the Kruskal-Wallis test.

[0030] Figure 5A It is a graph showing the pre-treatment prediction of the time to develop severe irAE episodes in patients treated with combination therapy. The cut point was optimized using a composite model score (trained with LOOCV). Only patients (n=23) who did not experience early progression from batch cohorts 1 and 2 were analyzed. Statistical significance was assessed by a two-sided log-rank test.

[0031] Figure 5B is a set of two graphs showing TCR clonality dynamics associated with the development of severe irAEs in patients treated with combination therapy. Left: Change in TCR clonality from baseline after initiation of combination therapy (measured by 1-Pielou uniformity), with colors indicating future irAE status. Right: Same as left, but showing changes in clonality according to future irAE status. Significance was determined by two-sided unpaired Wilcoxon rank sum test.

[0032] Figure 5C Is displayed from Figure 5BFigure 3. Enrichment of CD4 T 5+3 gene signatures in CD4 T cells from pre-treatment PBMC samples from 3 patients analyzed in the study, all of whom developed severe irAEs and showed TCR clonal expansion after ICI initiation ( Fig.15D ). The box centerline, box borders, and whiskers indicate the median, first and third quartiles, and minimum and maximum values, respectively. The points represent cells analyzed by scRNA-seq and annotated by Azimuth (naive CD4, n=245 cells; CD4 TCM, n=320 cells) or by their clonal persistence from baseline to early on-treatment time points (persistent CD4, n=190 cells). The most persistent CD4 clonotype in this analysis showed evidence of clonal expansion ( Fig.15F and G). Significance relative to persistent cells was determined by two-sided unpaired Wilcoxon rank sum test. ssGSEA denotes single sample GSEA.

[0033] Figure 5D Figure 2 is a plot of the difference from no severe irAEs stratified by the extent of TCR clonal expansion after starting combination therapy, measured by the change in 1-Pielou uniformity. Patients were divided into the following three categories: no clonal expansion (n=5), moderate (n=5), and high clonal expansion (n=5). Statistical significance was assessed by a two-sided log-rank test.

[0034] Figure 6 Color-coded graph showing large-scale assessment of circulating leukocytes in autoimmune diseases. Enrichment of circulating leukocyte levels in two autoimmune diseases relative to healthy controls. Leukocyte composition determined by CIBERSORTx. Significance determined by two-sided unpaired Wilcoxon rank sum test and comprehensive meta z-score. Fig.16 Details of the analysis workflow and underlying datasets are provided.

[0035] Fig. 7A Figure 3 is a UMAP representation of pre-treatment peripheral blood leukocytes analyzed by droplet-based scRNA-seq (10xGenomics) from 13 patients with metastatic melanoma, colored by major cell lineage, severe irAE status, TCR expression by scV(D)J-seq, and BCR expression by scV(D)J-seq (with Figure 3A related).

[0036] Figure 7B Schematic representation of unsupervised hierarchical clustering (average linkage) of the mean log2 transcriptome of each CD4 T cell cluster identified from scRNA-seq data.

[0037] Figure 7Care dot plots showing the average expression of key activation (HLA-DX, MKI67) and lineage markers (SELL, CCR7) in CD4 T cell clusters.

[0038] Fig.7D is a plot of unsupervised hierarchical clustering (average linkage) of the mean log2 transcriptome of each CD4 T cell cluster identified from the scRNA-seq data, showing all pairwise combinations of scRNA-seq clusters in each major cell type analyzed (B cells, CD4 T cells, CD8 T cells, NK cells, monocytes). Of the 82 possible pairwise combinations, CD4 T5+3 achieved the highest Spearman correlation and the strongest association with severe irAEs against CD4 TEM calculated by CyTOF. Cells annotated as "T / NKT" are collapsed to CD8 T cells.

[0039] Fig. 7E Figure 2 is an unsupervised hierarchical clustering (average linkage) of the mean log2 transcriptome for each CD4 T cell cluster identified from the scRNA-seq data, showing all pairwise combinations ranked by the mean of each feature after normalization to unit variance (mean = 0 and standard deviation = 1). In this analysis, the –log10 P values ​​(two-sided unpaired Wilcoxon rank sum test) for association with severe irAEs were normalized to unit variance, regardless of the direction of association.

[0040] Fig. 8A Shown is a UMAP map of the scRNA-seq data generated in this work, embedded and labeled by Azimuth, using a reference PBMC atlas of 162,000 cells (analyzed by scRNA-seq and 228 antibodies).

[0041] Figure 8B is a confusion matrix showing the phenotypic markers determined by marker genes and unsupervised clustering (rows; and Figure 3A and Fig. 7A The consistency between the reference-guided annotations (columns) compared with Azimuth was shown in Figure 3. A total of 85% of single cells assigned to major lineage groups (B cells, CD4 T, CD8 T, NK cells, monocytes) by Azimuth were assigned to the same identity by canonical marker gene assessment. Given the absence of NKT cells in the reference atlas used for Azimuth, the T / NKT cluster defined by the unsupervised analysis was relabeled as CD8 T cells.

[0042] Figure 8C is with Figure 3BFigure of the same analysis, but showing all 27 phenotypic states identified by Azimuth. Among these states, CD4 TEM was most associated with severe irAEs and CD4TEM counted by CyTOF. The group combining CD4 TEM and CD4 proliferation status was also strongly associated with severe irAEs. The latter showed the highest expression of HLA-DX and the lowest expression of SELL (Figure d), consistent with the activated CD4 TEM phenotype.

[0043] Fig.8D is a dot plot depicting key activation and lineage markers in CD4 T cell states annotated by Azimuth.

[0044] Fig. 8E is a set of violin plots showing protein expression levels derived from Azimuth using antibody-derived tag (ADT) data, which supports Figure 8C A and F show the combination of CD4 TEM and CD4 proliferation status.

[0045] Fig.8F Grid showing the performance of the top ranked cell subsets identified by Azimuth and unsupervised clustering for prediction of severe irAEs. The combined CD4 T 5+3 clustering ( Figure 3B ) than the highest ranked reference guide group ( Figure 3C ) were more associated with severe irAEs and CyTOF. Statistical significance was calculated using a two-sided unpaired Wilcoxon rank sum test. The data in all figures shown are from the 13 samples analyzed by scRNA-seq in Figure 3.

[0046] Fig.9A was shown in 13 PBMC samples analyzed by scRNA-seq ( Figure 1 and 3A ) Association between the occurrence of severe irAEs and pre-treatment levels of T cell states identified by unsupervised clustering (left) and with memory-like T cell state levels identified by Azimuth (right). Activated cells were defined as cells expressing HLA-DX or MKI67 (CPM>0); resting cells were defined by the absence of HLA-DX and MKI67 expression (CPM=0).

[0047] Fig. 9B A set of graphs showing the analysis of activated, resting, and parental T cell subsets associated with the development of severe irAEs. Left: For all 18 patients analyzed in the single-cell discovery cohort ( Figure 1 and 2A), association between the occurrence of severe irAEs and pretreatment levels of memory T cell subsets, total CD4 and CD8 T cells, and total T cells quantified by CyTOF. The activated phenotype was defined as CD38+ or HLA-DR+ or Ki67+. The resting phenotype was defined as CD38–HLA-DR–Ki67–. Right: ROC plot showing the performance of activated and resting CD4 TEM subsets (left) for predicting the occurrence of severe irAEs. Cell fractions were assessed relative to total PBMC content. Statistical significance in a, b was determined by a two-sided unpaired Wilcoxon rank sum test, and nominal -log10 P values ​​are shown. For associations without severe irAEs, the -log10 P values ​​were further multiplied by -1.

[0048] Fig. 10A Schematic diagram showing key TCR diversity measures and the impact of cell abundance, TCR richness, and different clonal repertoires on these measures. Hypothetical CD4 naive and TEM cell subsets are shown as examples. Triangles depicting size differences are not drawn to scale.

[0049] Fig. 10B is a plot of the mean Shannon entropy versus the mean clonality (1-Pielou uniformity) for each CD4 T cell state (identified by unsupervised clustering of scRNA-seq data). Figure 3B and C) is a TEM state enriched for activated cells, as expected from this phenotype, the CD4 T 5+3 displays increased clonality relative to the other CD4 states, while also displaying higher diversity (Shannon entropy), indicating increased abundance.

[0050] Fig. 10C is an EM-like CD4 T cell state with available scTCR clonotype data (from Figure 3F )'s distribution diagram.

[0051] Fig. 10D Figure 2 is a set of graphs showing the association between TCR diversity (Shannon entropy) and the occurrence of severe irAEs in pseudo-bulk T cells from pre-treatment blood, showing the presence of irAEs identified by scRNA-seq (left) and after removal of Fig. 10C All T cell states of EM-like state are shown (no severe irAE, n=5 patients; with severe irAE, n=4 patients). The borders and whiskers of the boxes represent the median, first and third quartiles, and minimum and maximum values, respectively.

[0052] Fig.10E Is displayed with Fig. 10DThe same association plot as in , but only for the EM-like state is shown. The borders and whiskers of the box indicate the median, first and third quartiles, and minimum and maximum values, respectively.

[0053] Fig.10F It is a graph showing the area under the curve (AUC) of the association between peripheral TCR diversity (Shannon entropy) before treatment and the occurrence of severe irAEs, showing all combinations of constituent cell states in e, including CD4T 5+3 clusters after being restricted to activated cells (CPM>0 of HLA-DX or MKI67). It is worth noting that in this analysis, no other combination of activated EM-like states reached AUC>0.85.

[0054] Figure 10G is a graph showing BCR clonotype diversity (Shannon entropy), showing each B cell state identified by unsupervised clustering ( Figure 3A ).exist Fig. 10B In DF and DF, only patients with at least 100 TCR clones were analyzed (n=9). Figure 10G The agreement of the same patients was analyzed in The box boundaries and whiskers indicate the median, first and third quartiles, and minimum and maximum values, respectively.

[0055] Fig.11A is a graph showing the expression of developmentally regulated marker genes in major CD4 T cell subsets from the LM22 signature matrix (MAS5 normalized), showing the LM22 reference signature of activated CD4 memory T cells with TEM spectra.

[0056] Fig. 11B is a graph showing CIBERSORTx versus mass cytometry for enumeration of activated CD4 memory T cells in pre-treatment peripheral blood of 17 patients with metastatic melanoma. Linear regression lines with 95% confidence intervals are shown. Agreement and significance were determined by Pearson's r and two-sided t-tests, respectively. Although activated CD4 memory T cells quantified by CyTOF are defined by CD38 expression in this figure, other activated CD4 TEM subsets were also significantly correlated with CIBERSORTx ( Fig. 11C ).

[0057] Fig. 11C is a cross-correlation plot of lymphocyte subset frequencies determined by CyTOF and CIBERSORTx. Act indicates activated.

[0058] Fig.11Dis a cross-correlation plot showing the correlation between activated CD4 memory T cell levels inferred by CIBERSORTx and 14 memory T cell states analyzed by CyTOF, including manually gated CD38+ activated subsets within each population in PBMCs from 17 metastatic melanoma patients.

[0059] Fig.11E Scatter plots depicting the global correlation of lymphocyte subsets counted by CIBERSORTx and flow cytometry in peripheral blood samples from five healthy subjects analyzed by bulk RNA-seq. Linear regression lines with 95% confidence intervals are shown. Agreement and significance were determined by Pearson’s r and two-sided t-tests, respectively. Because cytometry underestimates monocytes to varying degrees compared with complete blood counts, all results in b–e are expressed as a function of total lymphocytes.

[0060] Fig.11F is a graph showing the distribution of activated CD4 memory T cell levels quantified by CyTOF (CD38+, HLA-DR+, or Ki67+CD4 TEM cells, n=28 patients), scRNA-seq (HLA-DX+ or MKI67+ cells in CD4 T clusters 5 and 3, n=13 patients), and CIBERSORTx (n=60 patients) in all irAE evaluable samples analyzed by each modality in this work. The center line of the box, the border of the box, and the whiskers represent the median, the first and third quartiles, and the minimum and maximum values, respectively. Statistical significance was determined by the Kruskal-Wallis test. ns indicates not significant (P>0.05).

[0061] Fig. 12A is a graph showing the association between baseline batch TCR diversity and the highest irAE grade observed for each patient in batch cohorts 1 and 2, showing Shannon entropy and stratified by therapy type. Patients treated with combination therapy were stratified by the following future irAE status: no severe irAE (n = 10) vs. severe irAE (n = 14 patients) (left) and irAE grade (right): 0 / 1 (n = 3), 2 (n = 7), 3 (n = 12), and 4 (n = 2). The two-group comparisons were evaluated by a two-sided unpaired Wilcoxon rank sum test. ns indicates not significant (P>0.05). Linear regression was applied to evaluate the median of each measurement grouped by irAE grade (inset). The significance of linear consistency was determined by a two-sided t-test. Grades 0 and 1 reflect no toxicity and asymptomatic toxicity, respectively, and are combined. The box center line, box boundaries and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively, within 1.5×IQR (interquartile range) of the box limits.

[0062] Fig. 12B is a graph showing the association between baseline batch TCR diversity and the highest irAE grade observed for each patient in batch cohorts 1 and 2, showing the Gini-Simpson index and stratified by therapy type. Patients treated with combination therapy were stratified by future EirAE status: no severe irAE (n = 10) vs. severe irAE (n = 14 patients) (left) and irAE grade (right): 0 / 1 (n = 3), 2 (n = 7), 3 (n = 12), and 4 (n = 2). The two-group comparisons were assessed by a two-sided unpaired Wilcoxon rank sum test. ns indicates not significant (P>0.05). Linear regression was applied to evaluate the median of each measurement grouped by irAE grade (inset). The significance of linear consistency was determined by a two-sided t-test. Grades 0 and 1 reflect no toxicity and asymptomatic toxicity, respectively, and are combined. The box center line, box boundaries and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively, within 1.5×IQR (interquartile range) of the box limits.

[0063] Fig. 12C is a graph showing the association between baseline batch TCR diversity and the highest irAE grade observed for each patient in batch cohorts 1 and 2, showing Shannon entropy and stratified by therapy type. Patients treated with PD1 monotherapy were stratified by future irAE status: no severe irAE (n = 26) vs. severe irAE (n = 3 patients) (left) and irAE grade (right): 0 / 1 (n = 19), 2 (n = 7), 3 (n = 2), and 4 (n = 1). Two-group comparisons were assessed by a two-sided unpaired Wilcoxon rank sum test. ns indicates not significant (P>0.05). Linear regression was applied to evaluate the median of each measurement grouped by irAE grade (inset). The significance of linear consistency was determined by a two-sided t-test. Grades 0 and 1 reflect no toxicity and asymptomatic toxicity, respectively, and are combined. The box center line, box boundaries and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively, within 1.5×IQR (interquartile range) of the box limits.

[0064] Fig.12Dis a graph showing the association between baseline batch TCR diversity and the highest irAE grade observed for each patient in batch cohorts 1 and 2, showing the Gini-Simpson index and stratified by therapy type. The two-group comparison was assessed by a two-sided unpaired Wilcoxon rank sum test. ns indicates not significant (P>0.05). Linear regression was applied to assess the median of each measurement grouped by irAE grade (inset). The significance of linear consistency was determined by a two-sided t-test. Grades 0 and 1 reflect no toxicity and asymptomatic toxicity, respectively, and are combined. The center line of the box, the border of the box, and the whiskers represent the median, the first quartile and the third quartile, and the minimum and maximum values ​​within 1.5×IQR (interquartile range) of the box limit, respectively.

[0065] Fig.13A is similar to Figure 4D but applied to two bulk cohorts (n=53 patients) using leave-one-out cross-validation (LOOCV).

[0066] Fig. 13B is similar to Figure 4C , but showing the model scores determined by LOOCV.

[0067] Fig. 13C is a graph showing the performance of the composite model compared to other candidate pre-treatment factors for predicting the occurrence of severe irAEs. As shown in the figure, the composite model was trained on batch cohort 1 (BC1) and validated on batch cohort 2 (BC2), and vice versa.

[0068] Fig.13D is a graph showing the performance of the composite model trained on bulk cohort 1 for predicting severe irAEs in different patient subgroups from bulk cohort 2. DCB indicates durable clinical benefit; NDB indicates no durable clinical benefit; GI indicates gastrointestinal.

[0069] Fig.13E is a graph showing the composite model score determined by LOOCV for all bulk cohort patients (n=24) treated with combination therapy, stratified by future irAE level: 0 / 1 (n=3), 2 (n=7), 3 (n=12), and 4 (n=2). The center line, the border of the box, and the whiskers represent the median, the first quartile, and the third quartile, and the minimum and maximum values ​​within 1.5×IQR (interquartile range) of the box limit, respectively. Statistical significance was determined by the Kruskal-Wallis test.

[0070] Fig.13F Is displayed using Fig.13EFigure 3 shows the performance of the model using scores in predicting the occurrence of grade 2+, 3+, or 4 irAEs in patients on combination therapy.

[0071] Figure 13G is a graph showing the composite model score determined by LOOCV versus the number of symptomatic irAEs (grade 2+) per patient in two bulk cohorts (n = 53 patients). The center line, box borders, and whiskers represent the median, first and third quartiles, and minimum and maximum values ​​within 1.5 × IQR (interquartile range) of the box limits, respectively. Statistical significance was determined by the Kruskal-Wallis test.

[0072] Fig.13H is a graph showing the composite model score determined by LOOCV versus the number of organ system toxicities per patient in two bulk cohorts (n=53 patients). The center line, box borders, and whiskers represent the median, first and third quartiles, and minimum and maximum values ​​within 1.5×IQR (interquartile range) of the box limits, respectively. Statistical significance was determined by the Kruskal-Wallis test.

[0073] Fig.13I It is a graph showing the distribution of irAEs in patients and organ systems. Patients from batch cohorts 1 and 2 were organized by reducing composite model scores (determined by LOOCV). The line that distinguishes high / low scores was optimized using LOOCV.

[0074] Fig.13J is a graph showing the fraction of patients in two bulk cohorts who developed irAEs in at least 2 organ systems compared to the fraction of patients in two bulk cohorts who did not develop irAEs, by Fig.13I The thresholds in stratified. Significance was determined by two-sided Fisher's exact test.

[0075] FIG. 14 is a graph showing the performance of the composite model for predicting the time to develop severe irAEs in validation batch cohort 2.

[0076] Fig.14AFigures (a–c) show Kaplan–Meier analyses of patients treated with combination or PD1 immune checkpoint blockade (a), combination therapy (b), or PD1 monotherapy (c) who did not experience severe irAEs in bulk cohort 2, stratified by composite model score. Statistical significance was calculated by a two-sided log-rank test. Training was performed on bulk cohort 1 in all assay groups, and the cut points for predicting severe irAEs were optimized for bulk cohort 1 using Youden’s J statistic. Of note, the analyses in a–c were landmarked between the start of treatment and three months after the start of treatment, as all severe irAEs occurred within this time period. Kaplan–Meier plots are shown for four months because of the extended follow-up of patients who did not experience any severe irAEs.

[0077] Fig.14A Figure 2 is a plot showing the Kaplan-Meier analysis of the absence of severe irAEs in patients treated with combination or PD1 immune checkpoint blockade in bulk cohort 2, stratified by composite model score. Statistical significance was calculated by two-sided log-rank test. Training was performed on bulk cohort 1 in all test groups, and the cut points for predicting severe irAEs were optimized for bulk cohort 1 using Youden's J statistic. Of note, these analyses were landmarked between the start of treatment and three months after the start of treatment, as all severe irAEs occurred during this time period. Kaplan-Meier plots are shown for four months because of the extended follow-up of patients who did not develop any severe irAEs.

[0078] Fig. 14B Figure 2 is a plot showing the Kaplan-Meier analysis of the absence of severe irAEs in patients treated with combination therapy in bulk cohort 2, stratified by composite model score. Statistical significance was calculated using a two-sided log-rank test. In all test groups, training was performed on bulk cohort 1, and the cut-points for predicting severe irAEs were optimized for bulk cohort 1 using Youden's J statistic. Of note, Fig.14A -Analysis in C has a milestone between the start of treatment and three months after the start of treatment because all severe irAEs occurred during this time. Kaplan-Meier plots are shown for four months because of the extended follow-up of patients who did not develop any severe irAEs.

[0079] Fig. 14CFigure 2 is a plot showing Kaplan–Meier analysis of patients treated with PD1 monotherapy without severe irAEs in bulk cohort 2, stratified by composite model score. Statistical significance was calculated by a two-sided log-rank test. Training was performed in bulk cohort 1 in all test groups, and the cut points for predicting severe irAEs were optimized for bulk cohort 1 using Youden’s J statistic. Of note, the analyses in a–c were landmarked between the start of treatment and three months after the start of treatment, as all severe irAEs occurred within this time period. The Kaplan–Meier plot is shown for four months because of the extended follow-up of patients who did not develop any severe irAEs.

[0080] Fig.15A Figure 2 shows the RNA-seq profiles from paired pre-treatment PBMC samples (n = 15 patients on combination therapy) by MiXCR (bulk RNA-seq) and Plot of the evenness (Pielou index) of the (genomic DNA) assembled TCR repertoire. Consistency and significance were determined by Spearman's rho and two-sided t-test, respectively.

[0081] Fig. 15B is with Figure 5B Similar graph, but showing clonality of each pre-treatment and on-treatment PBMC sample. Statistical significance was determined by two-sided paired Wilcoxon rank sum test. Ns indicates not significant (P>0.05).

[0082] Fig. 15C Figure 2 is a graph showing the fraction of pre-treatment peripheral blood TCR clonotypes detected during treatment in 15 patients on combination therapy, stratified by no severe (n = 6) and severe (n = 9) irAE status. Clonotypes with matching productive CDR3β chain nucleotide sequences were considered identical. The center line, box borders, and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively. Significance was determined by a two-sided unpaired Wilcoxon rank sum test.

[0083] Fig.15D Is displayed by Schematic representation of the identified persister T cell clones cross-referenced with scTCR-seq and scRNA-seq data from pre-treatment PBMCs from the same three patients (YUALOE, YUNANCY, YUHONEY), all of whom received combination therapy and developed severe ICI-induced toxicity.

[0084] Fig.15E Figure 2 is a dot plot showing the log2 expression of key lineage and major T cell state neutralization and activation markers annotated by Azimuth and persisting clones classified as CD4 and CD8 T cells.

[0085] Fig.15F is a graph showing the overall change from baseline in the frequency of persistent clonotypes, stratified by lineage (n=2 cell types) and patient (n=3). Data calculations yielded the sum of differences in frequency (% during treatment - % before treatment). Bars represent mean + / - SD.

[0086] Figure 15G For display A set of graphs for peripheral blood TCR-β analysis. Top: Changes in bulk TCR clonality from baseline (Figure 5b). Bottom: Changes in bulk TCR clonality from baseline (Figure 5b). Fig.15F Same as, but showing potential clonotypes, where circle size corresponds to pre-treatment clonal frequency Proportional.

[0087] Fig.15H is similar to Figure 5D but limited to blood draws on day 1 of the first cycle of combination therapy and <1 month thereafter (n=7 patients).

[0088] Fig.16 Schematic diagram of large-scale assessment of peripheral blood leukocytes in autoimmune diseases versus healthy controls. Schematic diagram describing the workflow and statistical meta-analysis to assess the enrichment of individual circulating leukocyte subsets in autoimmune diseases relative to healthy controls ( Figure 6 ). Briefly, CIBERSORTx was used to count 15 leukocyte subsets in bulk RNA-seq or microarray analysis of peripheral blood samples from patients with systemic lupus erythematosus 57-59 (SLE; n = 239) or inflammatory bowel disease (IBD; n = 348) compared to healthy controls. For each dataset and cell subset, a two-sided unpaired Wilcoxon rank sum test was applied to assess differences in relative abundance between healthy and disease phenotypes. Results were then combined by a meta-z statistical study (Meth).

[0089] Fig.17A Schematic diagram showing the gating hierarchy and staining results of CD4 T cell subsets and NKT cells analyzed by CyTOF from pre-treatment PBMCs (a–e). All CD4 T cell subsets, except T regulatory cells (Treg), were similarly gated as CD8 T cells. TCM indicates central memory T cells; TEM indicates effector memory T cells; EMRA indicates CD45RA+ terminally differentiated effector memory T cells.

[0090] Fig. 17B is a schematic diagram showing the gating hierarchy and staining results of activated CD4 TEM versus resting CD4 TEM cells analyzed by CyTOF from pre-treatment PBMCs.

[0091] Fig. 17C is a schematic diagram showing the gating hierarchy and staining results of monocyte subsets analyzed by CyTOF from pre-treatment PBMCs.

[0092] Fig.17D is a schematic diagram showing the gating hierarchy and staining results of B cell subsets analyzed by CyTOF from pre-treatment PBMCs.

[0093] Fig.17E is a schematic diagram showing the gating hierarchy and staining results of NK cell subsets analyzed by CyTOF from pre-treatment PBMCs.

[0094] FIG. 18 is a set of graphs showing a comparison of automated and manual cell state quantification from CyTOF data.

[0095] Fig.18A is a scatter plot showing the agreement between the frequencies of automatic gating (Astrolabe) and manual gating for the indicated peripheral blood cell types. Agreement was assessed by Pearson correlation and linear regression (95% confidence intervals are shown). Statistical significance was assessed using a two-sided t-test. Data from Figure 1 Patients analyzed by CyTOF (n=18).

[0096] Fig.18B is similar to Fig.18A Scatter plot as seen in , but for CD4 TEM cells. Fig. 7A Representative gating schemes for CD4 TEM are provided. Consistency was assessed by Pearson correlation and linear regression (95% confidence intervals are shown). Statistical significance was assessed using a two-sided t-test. Data from Figure 1 Patients analyzed by CyTOF (n=18).

[0097] Fig.18C Figure 2 is a graph showing the association between pre-treatment CD4 TEM abundance and the occurrence of severe irAEs when expressed as a fraction of total PBMCs, total T cells, or CD4 T cells. The box center line, box borders, and whiskers represent the median, first and third quartiles, and minimum and maximum values, respectively. Data from Figure 1 Patients analyzed by CyTOF (n=18).

[0098] Fig.19 is a schematic diagram representing the method described in the present invention. DETAILED DESCRIPTION

[0099] About 60% of melanoma patients treated with combination immune checkpoint inhibitors (ICIs) experience severe immune-related adverse events (irAEs), leading to treatment-related morbidity and mortality. However, there is no reliable method to predict the occurrence or timing of severe irAEs.

[0100] Pre- and on-treatment analysis of cell states and T-cell receptors predicts the onset and timing of immunotherapy toxicity. We specifically combined the abundance of activated CD4 T effector memory cells and the diversity of the T-cell receptor repertoire in peripheral blood to generate a composite biomarker predicting immunotherapy toxicity. Clonal expansion from pre- to on-treatment predicted the timing of severe toxicity. Targeted RNA sequencing panels allowed this analysis to be performed in a practical and cost-effective manner.

[0101] Immunotherapy toxicities (immune-related adverse reactions) can be severe, dangerous, life-threatening, and fatal. In practice, we cannot reliably predict them early and before treatment. Reliable prediction would facilitate toxicity pre-planning, early intervention, and more personalized and precise immunotherapy administration.

[0102] In various aspects, a method for predicting immunotherapy toxicity in patients is disclosed. The disclosed method is based on the following discovery: two factors derived from the analysis of peripheral blood samples are closely related to the occurrence of severe immune-related adverse reactions (irAE), and the factors include activated CD4 memory T cell abundance levels and batch TCR diversity. In various aspects, a liquid biopsy method for predicting immunotherapy toxicity in patients is disclosed, and the method includes obtaining peripheral blood samples from subjects before receiving immunotherapy treatment. In various aspects, the method also includes the abundance of activated CD4 memory T cells and the diversity of T cell receptors (TCR) in quantitative peripheral blood samples. In some aspects, the method also includes determining a model index for predicting the possibility of severe irAR in patients, wherein the model index includes a combination of the abundance of activated CD4 memory T cells and the diversity of T cell receptors (TCR). The method also includes if the value of the model index exceeds a threshold value, the patient is classified as likely to have severe irAR. In some aspects, the method also includes predicting the severity of irAR based on the value of the model index, wherein the higher value of the model index predicts more severe irAR. The threshold value of the higher value of the model index can be determined empirically or with reference to known clinical standards.

[0103] In various other aspects, methods for predicting immunotherapy toxicity in patients based on the degree of TCR amplification are disclosed, where TCR amplification is defined herein as an increase in TCR diversity in the early stages of immunotherapy relative to TCR diversity before treatment. These methods include obtaining a first peripheral blood sample from a patient before starting immunotherapy, and obtaining a second peripheral blood sample in the early stages of administering immunotherapy to the patient. The method also includes obtaining a first TCR diversity from the first peripheral blood sample and obtaining a second TCR diversity from the second peripheral blood sample. These methods also include subtracting the first TCR diversity from the second TCR diversity to obtain the degree of TCR amplification. In some aspects, these methods also include classifying the patient as likely to have severe irAR if the degree of TCR amplification exceeds a threshold. In some aspects, these methods also include predicting the time of onset of severe irAR based on the degree of TCR amplification.

[0104] Molecular Engineering

[0105] The following definitions and methods are provided to better define the present invention and guide those of ordinary skill in the art to practice the present invention. Unless otherwise indicated, terms should be understood by those of ordinary skill in the relevant art according to conventional usage.

[0106] As used herein, the terms "heterologous DNA sequence," "exogenous DNA segment," or "heterologous nucleic acid" all refer to a sequence that is derived from a source that is foreign to a particular host cell, or, if from the same source, has been modified from its original form. Thus, a heterologous gene in a host cell includes a gene that is endogenous to the particular host cell but has been modified, for example, by the use of DNA shuffling. These terms also include non-naturally occurring multiple copies of a naturally occurring DNA sequence. Thus, these terms refer to a DNA segment that is foreign or heterologous to a cell, or a DNA fragment that is homologous to a cell but at a location within the host cell nucleic acid where the element is not commonly found. The exogenous DNA segment is expressed to produce an exogenous polypeptide. A "homologous" DNA sequence is a DNA sequence that is naturally associated with the host cell into which it is introduced.

[0107] Expression vectors, expression constructs, plasmids or recombinant DNA constructs are generally understood to refer to nucleic acids produced by human intervention (including recombinant means or direct chemical synthesis), using a series of specific nucleic acid elements that allow transcription or translation of a specific nucleic acid in, for example, a host cell. An expression vector can be part of a plasmid, a virus or a nucleic acid fragment. Typically, an expression vector can include a nucleic acid to be transcribed that is operably linked to a promoter.

[0108] "Promoter" is generally understood to be a nucleic acid control sequence that directs the transcription of a nucleic acid. An inducible promoter is generally understood to be a promoter that mediates the transcription of an operably linked gene in response to a specific stimulus. A promoter may include necessary nucleic acid sequences near the start site of transcription, such as, in the case of a polymerase II type promoter, a TATA element. A promoter may optionally include a distal enhancer or repressor element, which may be located up to several thousand base pairs from the transcription start site.

[0109] As used herein, "transcribable nucleic acid molecule" refers to any nucleic acid molecule that can be transcribed into an RNA molecule. Methods are known for introducing constructs into cells in such a way that transcribable nucleic acid molecules are transcribed into functional mRNA molecules that are translated and thus expressed as protein products. Constructs that can express antisense RNA molecules can also be constructed to inhibit the translation of a specific RNA molecule of interest. For the practice of the present invention, conventional compositions and methods for making and using constructs and host cells are well known to those skilled in the art (see, e.g., Sambrook and Russel (2006) Condensed Protocols from Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, ISBN-10: 0879697717; Ausubel et al., (2002) Short Protocols in Molecular Biology, 5th edition, Current Protocols, ISBN-10: 0471250929; Sambrook and Russel (2001) Molecular Cloning: A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, ISBN-10: 0879695773; Elhai, J. and Wolk, CP 1988. Methods in Enzymology 167, 747-754).

[0110] The "transcription start site" or "start site" is the position surrounding the first nucleotide that is part of a sequence that is transcribed, also defined as position +1. All other sequences of the gene and its control regions can be numbered relative to this start site. Downstream sequences (i.e., other protein coding sequences in the 3' direction) can be designated positive, while upstream sequences (primarily control regions in the 5' direction) are designated negative.

[0111] "Operably linked" or "functionally linked" preferably refers to the association of nucleic acid sequences on a single nucleic acid fragment, such that the function of one nucleic acid fragment is affected by the other nucleic acid fragment. For example, if the regulatory DNA sequence and the DNA sequence encoding RNA or polypeptide are positioned so that the regulatory DNA sequence affects the expression of the coding DNA sequence (i.e., the coding sequence or functional RNA is under the transcriptional control of the promoter), the regulatory DNA sequence is said to be "operably linked" or "associated" with the DNA sequence encoding RNA or polypeptide. The coding sequence can be linked to the regulatory sequence in the sense or antisense direction. The two nucleic acid molecules can be part of a single continuous nucleic acid molecule and can be adjacent. For example, if the promoter regulates or mediates the transcription of a gene of interest in a cell, the promoter is operably linked to the gene of interest.

[0112] "Construct" is generally understood to be any recombinant nucleic acid molecule, such as a plasmid, cosmid, virus, autonomously replicating nucleic acid molecule, bacteriophage, or linear or circular single-stranded or double-stranded DNA or RNA nucleic acid molecule, which is derived from any source and is capable of genome integration or autonomous replication, including a nucleic acid molecule in which one or more nucleic acid molecules have been operably linked.

[0113] Constructs of the present invention may contain a promoter, a transcribable nucleic acid molecule, and a 3' transcription termination nucleic acid molecule, wherein the promoter is operably linked to a transcribable nucleic acid molecule, and the transcribable nucleic acid molecule is operably linked to a 3' transcription termination nucleic acid molecule. Constructs may also include, but are not limited to, additional regulatory nucleic acid molecules from, for example, a 3' untranslated region (3'UTR). Constructs may include, but are not limited to, the 5' untranslated region (5'UTR) of an mRNA nucleic acid molecule, which may play an important role in translation initiation and may also be a gene component in an expression construct. These additional upstream and downstream regulatory nucleic acid molecules may be derived from natural or heterologous sources relative to other elements present on the promoter construct.

[0114] The term "transformation" refers to the transfer of a nucleic acid fragment into the genome of a host cell, resulting in genetically stable inheritance. Host cells containing the transformed nucleic acid fragment are called "transgenic" cells, and organisms containing transgenic cells are called "transgenic organisms."

[0115] "Transformed," "transgenic," and "recombinant" refer to a host cell or organism, such as a bacterium, cyanobacteria, animal, or plant, into which a heterologous nucleic acid molecule has been introduced. The nucleic acid molecule may be stably integrated into the genome, as is widely known and disclosed in the art (Sambrook 1989; Innis 1995; Gelfand 1995; Innis and Gelfand 1999). Known methods of PCR include, but are not limited to, methods using paired primers, nested primers, single specific primers, degenerate primers, gene specific primers, vector specific primers, partially mismatched primers, and the like. The term "untransformed" refers to a normal cell that has not been subjected to a transformation process.

[0116] "Wild type" refers to a virus or organism found in nature without any known mutations.

[0117] The design, generation and testing of variant nucleotides and encoded polypeptides having the above-mentioned desired percentage identity and retaining the desired activity of the expressed protein are within the technical scope of the art. For example, directed evolution and rapid isolation of mutants can be performed according to the methods described in the references, including but not limited to Link et al., (2007) Nature Reviews 5 (9), 680-688; Sanger et al., (1991) Gene 97 (1), 119-123; Ghadessy et al., (2001) Proc Natl Acad Sci USA 98 (8) 4552-4557. Therefore, those skilled in the art can generate a large number of nucleotide and / or polypeptide variants (e.g., having at least 95%-99% identity with the reference sequences described herein) and screen for the desired phenotype according to conventional methods in the art.

[0118] Nucleotide and / or amino acid sequence identity percentage (%) is understood to be the percentage of nucleotides or amino acid residues that are identical in the candidate sequence compared to the reference sequence when two sequences are compared. In order to determine the identity percentage, the sequences are compared, and if necessary, room is introduced to achieve the maximum sequence identity percentage. The sequence alignment program that determines the identity percentage is well known to those skilled in the art. Often publicly available computer software is used, such as BLAST, BLAST2, ALIGN2 or Megalign (DNASTAR) software to align sequences. Those skilled in the art can determine the appropriate parameters for measuring the comparison, including any algorithm that realizes the maximum comparison needs on the full length of the sequence compared. When sequences are aligned, the percentage of sequence identity of a given sequence A relative to, with, or for a given sequence B (or it can be expressed as a given sequence A having or comprising a percentage of sequence identity to, with, or for a given sequence B) can be calculated as follows: percentage of sequence identity=X / Y100, where X is the number of residues scored as identical matches by the sequence alignment program or algorithm alignment of A and B, and Y is the total number of residues in B. If the length of sequence A is not equal to the length of sequence B, the percentage of sequence identity of A relative to B will not be equal to the percentage of sequence identity of B relative to A.

[0119] In general, conservative substitutions can be made at any position as long as the desired activity is retained. So-called conservative exchanges can be made, in which the substituted amino acid has similar properties to the original amino acid, for example, Asp for Glu, Asn for Gln, Ile for Val, Ile for Leu, and Thr for Ser. For example, amino acids with similar properties can be aliphatic amino acids (e.g., glycine, alanine, valine, leucine, isoleucine); hydroxyl-containing or sulfur / seleno amino acids (e.g., serine, cysteine, selenocysteine, threonine, methionine); cyclic amino acids (e.g., proline); aromatic amino acids (e.g., phenylalanine, tyrosine, tryptophan); basic amino acids (e.g., histidine, lysine, arginine); or acidic amino acids and their amides (e.g., aspartic acid, glutamic acid, asparagine, glutamine). A deletion is the replacement of an amino acid by a direct bond. Locations of deletions include the termini of a polypeptide and the junctions between individual protein domains. An insertion is the introduction of an amino acid into a polypeptide chain, with the direct bond formally replaced by one or more amino acids. The amino acid sequence can be modulated with the aid of computer simulation programs known in the art, which can generate polypeptides with, for example, improved activity or altered regulation. Based on these artificially generated polypeptide sequences, the corresponding nucleic acid molecules encoding such modulated polypeptides can be synthesized in vitro using the specific codon usage of the desired host cell.

[0120] "Highly stringent hybridization conditions" are defined as hybridization at 65°C in 6×SSC buffer (i.e., 0.9 M sodium chloride and 0.09 M sodium citrate). Given these conditions, the melting temperature (T) of the DNA duplex between the two sequences can be calculated. m ) to determine whether a given set of sequences will hybridize. If the melting temperature of a particular duplex is below 65°C under the salt conditions of 6X SSC, then the two sequences will not hybridize. On the other hand, if the melting temperature is above 65°C under the same salt conditions, then the sequences will hybridize. In general, the melting temperature of any hybridizing DNA:DNA sequence can be determined using the following formula: m =81.5℃+16.6(log 10 [Na + ]) + 0.41 (G / C content fraction) – 0.63 (formamide%) – (600 / l). In addition, for every 1% decrease in nucleotide identity, the T of the DNA:DNA hybrid m A decrease of 1°C-1.5°C (see, e.g., Sambrook and Russel, 2006).

[0121] Host cells can be transformed using a variety of standard techniques known in the art (see, e.g., Sambrook and Russel (2006) Condensed Protocols from Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, ISBN-10: 0879697717; Ausubel et al., (2002) Short Protocols in Molecular Biology, 5th edition, Current Protocols, ISBN-10: 0471250929; Sambrook and Russel (2001) Molecular Cloning: A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, ISBN-10: 0879695773; Elhai, J. and Wolk, CP 1988. Methods in Enzymology 167, 747-754). These techniques include, but are not limited to, viral infection, calcium phosphate transfection, liposome-mediated transfection, microparticle-mediated delivery, receptor-mediated uptake, cell fusion, electroporation, etc. Transfected cells can be selected and propagated to provide recombinant host cells containing the expression vector stably integrated into the host cell genome.

[0122]

[0123]

[0124]

[0125]

[0126] Exemplary nucleic acids that can be introduced into host cells include, for example, DNA sequences or genes from another species, or even genes or sequences derived from or present in the same species but incorporated into recipient cells by genetic engineering methods. The term "exogenous" also means genes that are generally not present in transformed cells, or perhaps genes that are not present in the form, structure, etc. found in the transformed DNA segments or genes, or genes that generally exist and wish to express (e.g., overexpress) in a manner different from the natural expression pattern. Therefore, the term "exogenous" gene or DNA means any gene or DNA segment introduced into the recipient cell, regardless of whether similar genes already exist in such cells. The type of DNA included in the exogenous DNA can include the following DNA: DNA already present in the cell, DNA from another individual of the same type of organism, DNA from different organisms, or DNA produced externally, such as a DNA sequence containing the antisense information of a gene, or a DNA sequence encoding a synthetic or modified version of a gene.

[0127] Host strains produced according to the methods described herein can be evaluated by a number of methods known in the art (see, e.g., Studier (2005) Protein Expr Purif. 41(1), 207-234; Gellissen, ed. (2005) Production of Recombinant Proteins: Novel Microbial and Eukaryotic Expression Systems, Wiley-VCH, ISBN-10: 3527310363; Baneyx (2004) Protein Expression Technologies, Taylor & Francis, ISBN-10: 0954523253).

[0128] Methods for downregulating or silencing genes are known in the art. For example, antisense oligonucleotides, protein aptamers, nucleotide aptamers, and RNA interference (RNAi) (e.g., small interfering RNA (siRNA), short hairpin RNA (shRNA), and microRNA (miRNA)) can be used to downregulate or eliminate the activity of expressed proteins (see, e.g., Fanning and Symonds (2006) Handb Exp Pharmacol. 173, 289-303G, describing hammerhead ribozymes and small hairpin RNAs; Helene, C., et al. (1992) Ann. NY Acad. Sci. 660, 27-36; Maher (1992) Bioassays 14 (12): 807-15, describing targeted deoxyribonucleotide sequences; Lee et al., (2006) Curr Opin Chem Biol. 10, 1-8, describing aptamers; Reynolds et al., (2004) Nature Biotechnology 22(3), 326–330, describing RNAi; Pushparaj and Melendez (2006) Clinical and Experimental Pharmacology and Physiology 33(5-6), 504–510, describing RNAi; Dillon et al. (2005) Annual Review of Physiology 67, 147–173, describing RNAi; Dykxhoorn and Lieberman (2005) Annual Review of Medicine 56, 401–423, describing RNAi). RNAi molecules are commercially available from a variety of sources (e.g., Ambion, TX; Sigma Aldrich, MO; Invitrogen). Several siRNA molecule design programs using a variety of algorithms are known in the art (see, e.g., Cenix algorithm, Ambion; BLOCK-iT TM RNAiDesigner, Invitrogen; siRNA Whitehead Institute Design Tools, Bioinformatics & Research Computing). The traits that affect defining the optimal siRNA sequence include the following: G / C content at the ends of the siRNA, Tm of a specific internal domain of the siRNA, siRNA length, position of the target sequence within the CDS (coding region), and nucleotide content of the 3' overhang.

[0129] The definitions and methods described herein are provided to better define the present invention and guide those of ordinary skill in the art to practice the present invention. Unless otherwise indicated, terms are to be understood according to conventional usage by those of ordinary skill in the relevant art.

[0130] In some embodiments, the numbers of the components, properties (such as molecular weight, reaction conditions, etc.) of certain embodiments of the present invention used for description and claim protection should be understood to be modified by the term "about" in some examples. In some embodiments, the term "about" is used to represent the value of the standard deviation of the mean value of the device or method for determining the value. In some embodiments, the numerical parameters proposed in the written description and the attached claims are approximate values, which can be changed depending on the desired properties sought to be obtained by the specific embodiment. In some embodiments, the numerical parameters should be interpreted according to the number of reported significant figures and by applying ordinary rounding methods. Although the numerical ranges and parameters of the wide range of some embodiments of the present invention are approximate values, the numerical values ​​reported in the specific examples are reported as accurately as possible. The numerical values ​​presented in some embodiments of the present invention may contain some errors, which are necessarily caused by the standard deviation found in their respective test measurements. The description of the range of numerical values ​​herein is intended only as a shorthand method for citing each individual numerical value within the range individually. Unless otherwise specified herein, each individual value is incorporated into the specification as if it were individually described herein.

[0131] In some embodiments, unless otherwise specifically stated, the terms "a" and "an" and "the" and similar references used in the context of describing a particular embodiment (particularly in the context of certain claims below) may be interpreted to cover both the singular and the plural, unless specifically stated otherwise. In some embodiments, the term "or" as used herein (including the claims) is used to mean "and / or" unless explicitly indicated to refer to only alternatives or the alternatives are mutually exclusive.

[0132] The terms "comprise", "have" and "include" are open-ended linking verbs. Any form or tense of one or more of these verbs, such as "comprise", "comprises", "have", "have", "includes" and "includes", are also open-ended. For example, any method that "comprises", "has" or "includes" one or more steps is not limited to covering only these one or more steps, but may also cover other unlisted steps. Similarly, any composition or device that "comprises", "has" or "includes" one or more features is not limited to having only these one or more features, and may cover other unlisted features.

[0133] Unless otherwise specified herein or the context otherwise clearly contradicts, all methods described herein can be performed in any suitable order. The use of any and all examples or exemplary language (e.g., "such as") provided with respect to certain embodiments herein is intended only to better illustrate the present invention and does not limit the scope of the present invention. Any language in the specification should not be construed as representing any unclaimed element that is essential to the practice of the present invention.

[0134] The grouping of the alternative elements or embodiments of the present invention disclosed herein should not be construed as limiting. Each group member can be quoted and required individually, and can also be combined in any way with other members of the group or other elements found herein. For convenience or patentable reasons, one or more members in the group can be included in the group or deleted from the group. When any such inclusion or deletion occurs, this specification is considered to include the modified group, thereby satisfying the written description of all Markush groups used in the appended claims.

[0135] As will be understood based on the foregoing description, the above aspects of the present invention can be implemented using computer programming or engineering techniques, including computer software, firmware, hardware, or any combination or subset thereof. Any such resulting program with a computer readable code tool can be implemented or provided in one or more computer readable media, thereby manufacturing a computer program product, i.e., a product, according to the aspects discussed in the present invention. A computer readable medium can be, for example, but not limited to, a fixed (hard) drive, a floppy disk, an optical disk, a magnetic tape, a semiconductor memory (such as a read-only memory (ROM)), and / or any sending / receiving medium (such as the Internet or other communication networks or links). Products containing computer code can be manufactured and / or used by executing code directly from a medium, by copying code from a medium to another medium, or by transmitting code over a network.

[0136] These computer programs (also referred to as programs, software, software applications, "apps" or code) include machine instructions for a programmable processor and may be executed in high-level procedural and / or object-oriented programming languages ​​and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and / or "computer-readable medium" refer to any computer program product, apparatus and / or device (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. However, "machine-readable medium" and "computer-readable medium" do not include transient signals. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0137] As used herein, a processor may include any programmable system, including systems using microcontrollers, reduced instruction set circuits (RISC), application specific integrated circuits (ASICs), logic circuits, and any other circuits or processors capable of performing the functions described herein. The above examples are merely illustrative and are not intended to limit the definition and / or meaning of the term "processor" in any way.

[0138] As used herein, the terms "software" and "firmware" are interchangeable and include any computer program stored in a memory for execution by a processor, including RAM memory, ROM memory, EPROM memory, EEPROM memory, and non-volatile RAM (NVRAM) memory. The above memory types are examples only and thus do not limit the types of memory that can be used to store computer programs.

[0139] In one aspect, a computer program is provided and the program is mounted on a computer readable medium. In one aspect, the system is executed on a single computer system without requiring connection to a server computer. In another aspect, the system is (Windows is a registered trademark of Microsoft Corporation, Redmond, Washington)). On the other hand, the system operates in a host environment and (UNIX is a registered trademark of X / Open Company Limited, Reading, Berkshire, United Kingdom) server environment. The application is flexible and can run in a variety of different environments without compromising any key functionality.

[0140] In some aspects, the system includes multiple components distributed between multiple computing devices. One or more components can be in the form of computer executable instructions embodied in a computer readable medium. The system and processing are not limited to the specific aspects described herein. In addition, each system and each component of the processing can be practiced independently and separately from other components and processing described herein. Each component and processing can also be used in combination with other assembly packages and processing. This aspect can enhance the function and operation function of a computer and / or computer system.

[0141] All publications, patents, patent applications, and other references cited in this application are incorporated herein by reference in their entirety for all purposes to the same extent as if each individual publication, patent, patent application, or other reference was specifically and individually indicated to be incorporated herein by reference in its entirety. Citation of a reference herein should not be construed as an admission that it is prior art to the present invention.

[0142] Having described the invention in detail, it is apparent that modifications, variations and equivalent embodiments are possible without departing from the scope of the invention as defined in the appended claims. In addition, it should be understood that all examples in the present invention are provided as non-limiting examples.

[0143] Example

[0144] The following non-limiting examples are provided to further illustrate the present invention. It should be understood by those skilled in the art that the techniques disclosed in the following examples represent methods that the inventors have found to function well in the practice of the present invention and therefore can be considered to constitute embodiments of its practice mode. However, according to the present invention, it should be understood by those skilled in the art that many changes can be made in the disclosed specific embodiments without departing from the spirit and scope of the present invention, and still obtain similar or similar results.

[0145] Example 1: T cell characteristics associated with toxicity of immune checkpoint blockade in patients with melanoma

[0146] The following examples describe methods for predicting the likelihood of occurrence of serious immune-related adverse events (irAEs) associated with the administration of immunotherapy in melanoma patients.

[0147] summary:

[0148] Up to 60% of patients with melanoma treated with immune checkpoint inhibitors (ICIs) experience severe immune-related adverse events (irAEs). However, it is unclear whether there is a universal baseline immune status before the onset of irAEs. In this paper, time-of-flight mass cytometry, single-cell RNA sequencing, single-cell V(D)J sequencing, bulk RNA sequencing, and bulk T-cell receptor (TCR) sequencing were applied to study peripheral blood samples from melanoma patients treated with anti-PD-1 monotherapy or anti-PD-1 and anti-CTLA-4 combined ICI. By analyzing 93 blood samples before and during early ICI treatment and 3 patient cohorts (n=27, 26, and 18), 2 pre-treatment factors in the circulation (i.e., activated CD4 memory T cell abundance and TCR diversity) were found to be associated with the occurrence of severe irAEs, regardless of organ system involvement. We also explored the on-treatment changes in TCR clonality in patients receiving combination therapy, and these findings were associated with the severity and timing of irAE onset. These results demonstrate circulating T cell characteristics associated with ICI-induced toxicity, which has important implications for improved diagnosis and clinical management.

[0149] result:

[0150] In this study, we systematically evaluated immune signatures associated with ICI-induced toxicity in the peripheral blood of patients with metastatic melanoma. In different single-cell and bulk analysis modes, common T-cell signatures associated with the occurrence of severe irAEs within three months of treatment initiation were identified. These signatures were independent of key clinical variables, including durable clinical response and treatment with anti-PD-1 monotherapy or anti-PD-1 and anti-CTLA-4 combination therapy. Using these findings, predictive models for the occurrence of irAEs were developed and their use was explored for identification in pre-treatment and early treatment of ICI-induced toxicity.

[0151] Clinical cohort characteristics

[0152] To investigate candidate risk factors associated with severe (grade 3+) irAEs, 78 patients with metastatic melanoma were identified, of whom 71 were evaluable after application of exclusion criteria ( Figure 1 ). Of these patients, 33 were treated with anti-PD-1 monotherapy, 38 were treated with anti-PD-1 plus anti-CTLA-4 combination therapy, and 90% of patients had no history of ICI use. All patients were closely monitored for the occurrence of irAEs during and after ICI treatment (median follow-up time, 14.9 months; median time for grade 3+ irAEs, 1.5 months). Most patients experienced one or more irAEs, ranging from mild (grade 1) to life-threatening (grade 4) and affecting multiple organ systems, which were classified by board-certified clinicians according to standardized criteria (CTCAE v.5.0). The 71 patients were divided into the following 3 non-overlapping cohorts: a single-cell discovery cohort and a larger bulk cohort (divided into training and validation groups) ( Figure 1 ).

[0153] Determinants of severe irAEs from pretreatment blood.

[0154] Pre-treatment peripheral blood was analyzed to identify cellular determinants of severe irAEs using mass cytometry.

[0155] First, we performed a single-cell discovery cohort of 18 patients. Figure 1 and 2A ) were subjected to high-dimensional single-cell analysis of pre-treatment peripheral blood samples from 8 patients who experienced severe irAEs after the start of treatment. By applying time-of-flight mass cytometry (CyTOF) to analyze 35 leukocyte markers in each sample, 20 different subsets from nearly 800,000 evaluable cells were analyzed, covering 7 major monocytic lineages (B cells, plasmablasts, CD4 and CD8 T cells, natural killer (NK) cells, natural killer T (NKT) cells, monocytes) ( Figure 2B ,2C , 7, and 8A). Next, each subgroup was asked about severe irAE outcomes ( Figure 2C Among all subpopulations, only CD4 effector memory T (TEM) cells were significant after multiple hypothesis correction, and higher levels in pre-treatment blood were associated with severe irAEs (P = 0.0002; Q = 0.004; Figure 2C , 2D , 7B, 8B, and 8C). To confirm this finding, we examined the same peripheral blood samples from 13 patients using 5' droplet-based 10x Chromium single-cell RNA sequencing (scRNA-seq) paired with single-cell V(D)J sequencing (scV(D)J-seq) of TCR and B cell receptor (BCR) clonotypes. After quality control ( Fig. 7A ), the 5' assay yielded 24,807 cells and seven major lineages (based on classical marker gene expression classification ( Figure 3A )). Using unsupervised clustering, 32 different transcriptional states were identified in 7 cell types ( Figure 3A ). The correlation between cell state abundance and the occurrence of severe irAEs was then calculated. It is worth noting that among these 32 cell states, CD4 T cell state 5 (lacking expression of CCR7 and SELL (CD62L) and consistent with CD4 TEM cells) was found to be most closely associated with the occurrence of severe irAEs (nominal P = 0.05, two-sided unpaired Wilcoxon rank sum test; Figure 3B This status also correlated best with CD4 TEM levels measured by CyTOF ( Figure 3B When the joint probability of this outcome was considered by permutation test, the calculated empirical P value was 0.003. Further analysis showed that CD4 T cell state 3 (shown by unsupervised hierarchical clustering to be closely associated with state 5 ( Figure 7B )) also showed an expression profile consistent with CD4 TEM ( Figure 3C and 7C When combined with state 5, the resulting cluster (CD4 T 5+3) was more significantly associated with severe irAE occurrence and CD4 TEM levels (counted by CyTOF) ( Figure 3B Indeed, within each major cell type, CD4 T 5+3 had the highest Spearman correlation against CD4 TEM levels (as counted by CyTOF) among all 82 possible pairwise combinations of cell states and had the strongest association with the development of severe irAEs ( Fig.7D and 7E ).

[0156] Differential gene expression analysis of other CD4 T cell states showed that CD4 T 5 and 3 were enriched for markers of activated effector cells (including HLA-DRA, MKI67, TNFRSF4 (OX40), CCL5, and IL32), and were depleted for markers of TCM cells (SELL / CD62L) and markers of naive T cells (CCR7, TCF7) ( Figure 3C and 7C ). Reference-guided cell labeling using Seurat Azimuth confirmed that CD4 TEM cells were most associated with severe irAEs and were most similar to the CD4 T 5+3 population identified by de novo analysis (Figure 8). In addition, when the CD4 T 5+3 population was subdivided into activated and quiescent subgroups based on the expression of classical activation markers (HLA-DX, MKI67), the activated subgroup showed the strongest association with the occurrence of severe irAEs (P = 0.002, two-sided unpaired Wilcoxon rank sum test; Figure 3D and 9A This finding was verified using reference-guided annotation by Azimuth and CyTOF ( Fig.9A and 9B ), indicating that activated CD4 TEM cells preferentially underlie severe ICI toxicity.

[0157] Given this observation, one wondered whether pre-treatment TCR diversity in activated CD4 T cells might also be associated with severe ICI toxicity. Indeed, in patients who experienced severe irAEs, single-cell TCR clonotype diversity (Shannon entropy) of activated CD4 T 5+3 cells was elevated (area under the receiver operating characteristic curve (AUC) = 0.90, P = 0.05; Figure 3E This suggests that TCR richness (defined as the number of unique clonotypes in a sample and a key component of diversity metrics including Shannon entropy) masks loss of diversity due to clonal expansion when activated CD4 TEM cells are quantified relative to total peripheral blood mononuclear cells (PBMCs). Fig. 10A and B). In other words, TCR richness of activated CD4 TEM cells in total PBMCs in patients destined to develop severe irAEs underlies the overall increase in TCR diversity prior to treatment. Of note, the definition of clonotypic diversity that includes richness has ample precedent in the literature, including studies of circulating and tumor-infiltrating T cells, which provides a solid foundation for their application in this work.

[0158] Although this association between TCR diversity and severe irAE occurrence was attenuated or absent in other T cell subsets when all evaluable T cells were combined, a striking trend was observed between bulk TCR diversity and severe irAE occurrence in pretreatment samples (AUC = 0.80; Figure 3E ). Moreover, this association was mainly attributed to CD4 T cells with effector memory characteristics (low CCR7 and SELL) ( Figure 3F and 10C -F). In contrast, differences in peripheral blood BCR diversity associated with severe irAEs were less pronounced ( Figure 10G Collectively, these findings suggest an association between a more diverse TCR repertoire in CD4 TEM cells at baseline (broadly reflected in the bulk peripheral blood) and the progression of severe ICI toxicity.

[0159] Expanded analysis of T cell signatures associated with irAEs.

[0160] After identifying candidate pretreatment determinants of severe irAE development, these findings were validated in a larger independent patient group. Based on sample size estimates, bulk RNA sequencing (bulk RNA-seq) was applied to pretreatment peripheral blood samples from an additional 53 patients with metastatic melanoma across two cohorts (n = 26 and 27) treated with single-agent (anti-PD-1, n = 29) or combination (anti-PD-1 and anti-CTLA-4, n = 24) checkpoint blockade. Figure 1 ). To evaluate circulating immune signatures in bulk transcriptome analysis, CIBERSORTx, a machine learning method for counting cell subsets from bulk tissue expression profiles, and MiXCR, a computational method for V(D)J clonotype assembly and quantification from bulk RNA-seq data, were applied. Using peripheral blood (CyTOF) from 17 patients with melanoma, the accuracy of CIBERSORTx for deconvolution of major blood lineages, including specificity for activated CD4 memory T cell (TM) signatures against activated CD4 TEM cells, was confirmed by direct comparison with cytometry assays (Figure 11). Notably, of the 13 PBMC subsets evaluable by CIBERSORTx, only levels of activated CD4 TM cells were associated with severe irAEs ( Figure 4A ; P < 0.025 for combined and PD-1, HR = 8.3 and 14.8, respectively; Fig. 14B and C) or the cohort assessed by LOOCV (P = 0.0028 and HR = 12.2 for combination therapy, Figure 5A; P = 0.03 and HR = 9.0 for PD-1 therapy). The model also predicted the time to develop severe irAEs in the multivariate model (independent of therapy type, age, sex, and other key parameters).

[0161] Peripheral TCR clonal expansion associated with severe irAEs.

[0162] Previous case reports of patients with melanoma experiencing lethal ICI-mediated toxicity showed evidence of clonally expanded autologous or virus-reactive T cells in affected tissues, linking T cell clones that recognize both self and pathogens to lethal toxicity. Therefore, it was hypothesized that pretreatment TCR clonotypes in peripheral blood might show a greater propensity to expand in patients destined to develop irAEs after initiation of ICI therapy. To examine this, immunoSEQ was applied to analyze bulk TCR-β repertoires in paired pretreatment and early on-treatment PBMC samples collected from 15 patients with metastatic melanoma receiving combination therapy. Using a TCR clonality index that is robust to variations in the number of captured clones (Pielou uniformity), significant concordance between MiXCR (bulk RNA-seq) and immunoSEQ (DNA) was confirmed in pretreatment samples from these 15 patients, emphasizing the integrity of the composite model in bulk cohorts 1 and 2 ( Fig.15A ). TCR clonal expansion (ie, clonal dominance) was then assessed after treatment initiation, as measured by an increase in 1-Pielou evenness. Supporting this hypothesis, a significant increase in both TCR clonal expansion and persistence of the baseline clone was observed in patients who developed severe irAEs compared with those who did not develop severe irAEs ( Figure 5B , 15B , 15C). In patients with severe irAE who underwent scRNA-seq and single-cell TCR sequencing (scTCR-seq) (n=3), preferential expansion of the activated CD4 TEM compartment was observed among clones detected in both blood draws ( Fig.15D -G). In addition, persistent CD4 T cell clones were highly enriched for the CD4 T 5+3 population identified by scRNA-seq analysis ( Figure 5C and 15E ).

[0163] We also explored whether the extent of TCR clonal expansion during early treatment was associated with the time to severe irAEs. In fact, patients with greater TCR clonal expansion developed severe irAEs more quickly, whether assessed by the log-rank test or by tertiles of ranks through Cox proportional hazards regression (P = 0.003, log-rank test; P = 0.014, p = 0.003). Figure 5D). This was independent of the time between blood draws and when the analysis was restricted to on-treatment blood draws obtained within one month of cycle 1 ICI ( Fig.15H ), these results are significant.

[0164] Circulating Leukocytes in Autoimmune Disease

[0165] Finally, we asked whether baseline peripheral blood analysis of patients at risk for severe irAEs parallels clinical autoimmunity. To this end, CIBERSORTx was applied to examine 15 leukocyte subsets in bulk peripheral blood transcriptomes across 6 studies and 587 patients with systemic lupus erythematosus (SLE) or inflammatory bowel disease (IBD) (vs. 191 healthy controls). A meta-analysis framework was used to integrate P values ​​and pathology across studies ( Fig.16 ), found that circulating activated CD4 TM cells were most significantly associated with autoimmune diseases relative to healthy individuals ( Figure 6 These data suggest that severe irAEs may represent a subclinical or underlying autoimmune state that is not clinically masked at the time of ICI administration, which is consistent with recent case reports and multi-institutional data showing that autoimmune patients treated with immune checkpoint blockade are prone to experience flare-ups of autoimmune symptoms.

[0166] discuss

[0167] In this study, two baseline features (abundance of activated CD4 TM cells in peripheral blood and a more clonally diverse TCR repertoire) were identified as promising determinants of ICI-induced irAEs in patients with metastatic melanoma. Although previous studies have (1) linked activated T cells and clonally expanded TCRs in postmortem tissues to fatal irAEs (myocarditis, encephalitis) and (2) linked effector CD4 T cells to organ-specific irAEs (destructive thyroiditis, hepatitis), this work extends the scope of these findings to pre-treatment T cell properties of irAEs in multiple organ systems. Integrating these features into a composite model predicted a greater risk of severe irAEs and demonstrated sufficient granularity to distinguish between different irAE grades and burdens.

[0168] A significant correlation was also identified between early T-cell clonal expansion and the time to onset of severe irAEs in patients treated with combination therapy. Future studies are needed to further characterize this finding and to elucidate the relative contributions of CD4 and CD8 T cells to the clonal dynamics associated with irAEs.

[0169] Consistent with the possibility of a common immune mechanism underlying the development of irAEs and autoimmunity, elevated levels of activated CD4 TM cells were additionally observed in patients with SLE or IBD. Although it is reasonable to predict that patients with preexisting autoimmunity would be enriched for higher levels of activated CD4 T cells and higher rates of severe irAEs with ICI, no patients in the cohort had documented preexisting autoimmunity. Furthermore, such patients may have developed compensatory immune regulatory mechanisms prior to starting ICIs that alter their baseline irAE risk. However, it will be important to investigate this association in more detail in future studies and to determine whether circulating activated CD4 TM cells exhibit an increased propensity to recognize autoantigens in patients at risk for severe ICI toxicity. Indeed, patients with autoimmune diseases, particularly those with gastrointestinal or rheumatological conditions, who are treated with combination immunotherapy are at greater risk for flare-ups. More reliable identification of these at-risk patients during the ICI decision-making process could improve their outcomes.

[0170] This study has several limitations. First, this study adopted a retrospective design using banked clinical samples. Second, patients received anti-PD-1 monotherapy or anti-PD-1 plus anti-CTLA-4 combination therapy, which is associated with different risk analyses for the development of severe irAEs. Third, although most irAEs occurred within the first three months of ICI treatment initiation, some subgroups may occur later. Since the median time to development of severe irAEs in our cohort was 6.4 weeks (consistent with clinical trial data) and no irAEs occurred for more than 3 months, it is necessary to investigate whether these findings apply to late-onset irAEs. Fourth, the timing of intra-treatment peripheral blood collection during immunotherapy was not uniform with respect to treatment initiation. Finally, it is unclear whether these findings will apply to ICI-associated irAE risks in other cancer types.

[0171] Future studies should address these limitations, as well as more extensively apply single-cell analysis before and during early immunotherapy. In addition, it will be important to confirm our findings in larger multi-institutional cohorts and assess whether circulating immune determinants of ICI-induced toxicity differ based on the organ most likely involved. If prospectively validated, these findings could facilitate therapeutic adaptation to improve the risk profile of immune checkpoint blockade, thereby involving the prediction and potential prevention of ICI-mediated toxicity.

[0172] method

[0173] Study Design and Participants

[0174] In accordance with the Declaration of Helsinki (2013), samples analyzed in this study were collected with informed consent for research use as part of an observational registry study focused on melanoma and approved by the Institutional Review Boards of Yale University School of Medicine and Washington University School of Medicine. Eligible patients aged >18 years with metastatic melanoma were treated with ICI consisting of anti-PD-1 blockade (nivolumab or pembrolizumab) or combined immune checkpoint blockade (anti-PD-1 (nivolumab) and anti-CTLA-4 (ipilimumab); Figure 1 ). 90% of patients were lacking use of any prior immune checkpoint blockade at the time of pre-treatment blood draw. All patients underwent routine clinical assessment of irAEs and responses, performed by board-certified medical oncologists. Monitoring occurred before each cycle of ICI treatment (approximately every 3 weeks) and in some cases more frequently (e.g., inpatient staff monitoring of patients admitted with severe irAEs). The monitoring continued after the treatment course was completed, if applicable. All irAEs were classified according to the United States Health and Human Services Common Terminology Criteria for Adverse Events (CTCAE) v.5.0, with grades ≥2 and ≥3 considered symptomatic and severe, respectively. Within and across patient cohorts, irAEs affected multiple organ systems, including the gastrointestinal tract, skin, liver, pituitary, thyroid, adrenal, musculoskeletal, ocular, pancreatic, and cardiac systems ( Fig.13I ). Three patients (YUGIM, YUHERN, and YUTORY) experienced a systemic inflammatory syndrome associated with ICI administration. All severe irAEs occurred within three months of ICI initiation, a milestone period during which no patient in this cohort died. The response was scored as: durable clinical benefit, no durable benefit, or not evaluable (as described previously). Patients in the three cohorts who met the above eligibility criteria were identified, and pre-treatment PBMC samples were collected just before the first cycle of anti-PD-1 or combination ICI administration (median 0d; range 0-2 months). As Figure 1 As shown, PBMCs from each cohort were analyzed (pre-treatment for all patients and pre-treatment / on-treatment pairs for 15 patients).

[0175] Blood collection and processing.

[0176] Peripheral blood samples were collected in K2EDTA Vacutainer tubes (Becton Dickinson) and processed within 1 h of venous blood collection. PBMC extraction was performed by either the ammonium chloride or Lymphoprep (STEMCELL Technologies) protocol. The Lymphoprep protocol was applied according to the manufacturer's instructions. Following the ammonium chloride protocol, 4–8 ml of blood was mixed with 20 ml of cold ammonium chloride lysis buffer (0.1 M ammonium chloride, 0.01 M Tris-HCl) and incubated at room temperature for 5 min. The cells were then centrifuged at 300 g for 5 min and washed with 5 ml of cold PBS. PBMC samples were cryopreserved in 10% dimethyl sulfoxide / 90% FBS. Cryovials were placed in Nalgene Mr. Frosty containers (ThermoFisher Scientific) for 24 h and then stored in liquid nitrogen until cells and RNA were processed for expression analysis.

[0177] Mass cytometry.

[0178] Metal-conjugated antibodies were purchased either pre-conjugated from Fluidigm or purified from BioLegend, ThermoFisher Scientific, or Cell Signaling Technology and then conjugated to metals using the Maxpar Antibody Labeling Kit (Fluidigm) according to the manufacturer’s instructions. PBMCs were prepared for CyTOF from each of the 28 patients. Cryopreserved cell suspensions were first thawed by placing the cryovials in a 37°C water bath for 1–2 min without submerging the lids. Subsequently, 1–3 × 10 cells of the single-cell suspension were plated and thawed. 6 PBMCs were incubated with Human TruStain FcX (BioLegend) at room temperature for 10 minutes to block nonspecific antibody binding and then incubated with metal-conjugated antibodies against cell surface molecules on ice for 20 minutes. Cells were also incubated with Cell-ID Cisplatin (Fluidigm) according to the manufacturer's instructions to identify live cells. After treatment with intracellular fixation and permeabilization buffer (Thermo Fisher Scientific), cells were incubated with metal-conjugated antibodies against intracellular proteins. The cells were then washed and stained with Cell-ID Intercalator-Ir (Fluidigm) diluted in PBS containing 1.6% paraformaldehyde (Electron Microscopy Sciences) and stored at 4°C until collection. After the washing step, the Helios system (Fluidigm) was then used for <400s -1Sample collection is performed based on the event rate.

[0179] To reduce technical variation between samples, Ce beads were used in each sample and the data were normalized together using BeadNormalizer v0.3 (https: / / github.com / nolanlab / bead-normalization / wiki / Installing-the-Normalizer). To further reduce technical variation, sample processing and acquisition batches were limited to four batches, the same reagent batch was used for all samples, and no major adjustments were made to the Helios calibration. It is also noted that Astrolabe does not compare numerical intensities between samples; instead, Astrolabe analyzes each sample individually, assuming that a given subgroup is identical, regardless of potential marker intensity variations. Therefore, the platform is reported to be immune to batch effects.

[0180] Mass cytometry data analysis.

[0181] CyTOF data were initially analyzed by Cytobank v8.0 and v8.1 (Beckman Coulter), hierarchical clustering was optimized using the FlowSOM algorithm, and high-dimensional data were visualized using the viSNE algorithm (5000 iterations, perplexity = 100). Subsequent cell subset identification and data visualization were performed using the Astrolabe Cytometry Platform v3.6 and v4.0 (Astrolabe), which automatically labeled cell subsets using the Ek'Balam algorithm 71 (a knowledge-based hierarchical annotation strategy combined with unsupervised clustering). A total of 20 cell subsets across the main mononuclear lineages in peripheral blood were identified and quantified. For each patient sample, cell subset levels were normalized to a sum of 1, and cells that were unclassifiable based on protein marker expression were excluded from the analysis. To confirm Astrolabe, Cytobank was used to perform blind manual gating of major cell populations including CD4TEM cells (Figures 17, 18A, and 18B). The total abundance of CD4 TEM cells (either calculated as a fraction of total PBMCs or circulating T cells, but not as a fraction of CD4 T cells) was significantly associated with the development of severe irAEs ( Fig.18C ).

[0182] Flow cytometry.

[0183] PBMCs collected from five healthy donors were analyzed by flow cytometry ( Fig.11EBriefly, 2-5 million PBMC cells were treated with TruStain FcX Fc Receptor Blocking Solution (BioLegend) for 10 min at room temperature to block Fc receptors and then stained with fluorophore-labeled surface antibodies for 30 min at room temperature. The following antibodies were used to stain cells: FITC-conjugated anti-human CD45 (clone 2D1; BioLegend); AF700-conjugated anti-human CD3 (clone OKT3; BioLegend); APC-conjugated anti-human CD4 (clone OKT4; BioLegend); PE / Cy7-conjugated anti-human CD8 (clone SK1; BioLegend); APC-Cy7-conjugated anti-human CD19 (clone HIB19; BioLegend); PerCp / Cy5.5-conjugated anti-human CD14 (clone HCD14; BioLegend); and BV605-conjugated anti-human CD56 (clone 5.1H11; BioLegend). The cells were then washed twice with ice-cold PBS-based buffer (1x PBS, 2% FBS, 1 mM EDTA) and stained with 4',6-diamidino-2-phenylindole (DAPI) (BioLegend) to assess cell viability. Antibody capture beads (BD Biosciences) were used to compensate for each fluorophore in the experiment. The stained cells were analyzed by flow cytometry with the help of the operator using a MoFlo Legacy instrument (BeckmanCoulter) at the Siteman Flow Cytometry Center at the University of Washington School of Medicine. After excluding DAPI-positive cells and presumed doublets based on forward and side scatter analysis, major lymphocyte populations (including B cells, CD4 T cells, CD8 T cells, and NK cells) were counted as a percentage of total lymphocytes using FlowJo v.10 (FlowJo LLC).

[0184] scRNA-seq and scV(D)J-seq library preparation and sequencing.

[0185] Single cell suspensions were obtained from PBMC samples as described above and cell counts were performed using a hemacytometer (Thermo Fisher Scientific) or a Coulter counter (Beckman Coulter Life Sciences) according to the manufacturer's instructions to prepare 700-1200 viable cells / μl. -1The single cell suspension was then subjected to library preparation for scRNA-seq and paired scV(D)J-seq for TCR and BCR clonotypes using a 5' transcriptome kit (10x Genomics) according to the manufacturer's instructions. Complementary DNA libraries were sequenced on a NovaSeq instrument (Illumina) with 2 × 92 base pair (bp) paired-end reads, with a target of an average of 20,000 reads / cell.

[0186] scRNA-seq analysis (discovery cohort).

[0187] Raw scRNA-seq reads were barcoded and deduplicated using Cell Ranger v.3.1.0 and aligned to the hg38 reference genome to generate a sparse digital count matrix that was analyzed using Seurat v.3.1.5 or v.3.2.1 (ref. 72) to identify cell types and cell states. Outlier cells were identified and removed based on the following criteria: (1) mitochondrial content >25% or (2) cells expressing fewer than 100 or more than 1500-3000 genes, depending on sample level distribution. After normalization (NormalizeData) and variable feature identification (FindVariableFeatures with n=2000 features), FindIntegrationAnchors (dims=1:30) was applied to identify anchors and IntegrateData (with default parameters) for batch correction. After integration, principal component analysis (PCA) and uniform manifold approximation and projection (UMAP) were applied using the 2000 most variable genes and the first 30 principal components. FindClusters was applied to identify cell types and cell states, with the resolution parameter set to 3, resulting in 37 clusters.

[0188] All identified clusters were assigned to major cell lineages based on the expression of classical marker genes: CD3D / CD3Ehi=T cells; CD8A / CD8Bhi and NKG7 / GNLYlo=CD8 T cells; non-CD8 T cells with high IL7R expression and low NKG7 / GNLY=CD4 T cells; NKG7 / GNLYhi and CD3D / CD3Elo=NK cells; CD14 or FCGR3Ahi=monocytes; FCER1Ahi=dendritic cells (DC); MS4A1hi=B cells; HBBhi=erythrocytes; PPBPhi=platelets. Cells with high expression of CD3D / E and GNLY / NKG7 that were not annotated as CD8 / CD4 T cells were included in the T or NKT cell group, denoted as T / NKT. Clusters annotated as erythrocytes or platelets were omitted from further analysis. To assess the effective fold, cell barcodes were cross-referenced with single cell BCR (scBCR) and TCR (scTCR) clonotypes. By determining (1) the percentage of non-T cells that aberrantly mapped to a TCR clonotype (denoted as m) and (2) the frequency of annotated T cells with a matching scTCR clonotype (denoted as f) (i.e., the recovery rate), we calculated an effective fold ratio (m / f) of 2.2%. The same effective fold ratio was calculated for scBCR clonotypes mapped to non-B cells (also 2.2%). Since the effective fold ratio was quite low, all single cells with aberrant expression of TCR or BCR clonotype sequences were eliminated. PCA, UMAP, and FindClusters were then repeated as described above, resulting in 32 clusters. Two red blood cell clusters (marked by very high HBB expression) were retained and removed from the analysis, followed by a final round of PCA, UMAP, and FindClusters, resulting in the final set of 32 clusters (i.e., states) and Figure 3A and 7A The low-dimensional embedding shown.

[0189] All 32 states were evaluated for association with the occurrence of severe irAEs as measured by CyTOF ( Figure 3B The correlation between the abundance of CD4 TEM ( Figure 3B Among them, CD4 T cluster 5 has the strongest correlation with both variables ( Figure 3B). To determine the statistical significance of this result, the joint probability of (1) being ranked first by each measurement and (2) achieving a P value and Spearman correlation coefficient at least as strong as CD4 T cluster 5 was calculated. To calculate this probability empirically, a permutation scheme was implemented in which the cell fractions associated with each scRNA-seq cluster were independently shuffled across all patient samples and then evaluated for (1) and (2) above. By repeating this process 10,000 times, an empirical P value of 0.003 was calculated for CD4 T cluster 5. Pairwise combination analyses were also performed, restricting pairs of cell states to the same major cell type (B cells, CD4 T cells, CD8 T cells, NK cells, monocytes) to maintain biological consistency, and each of the 82 possible cell cluster combinations was compared with CD4 TEM levels (counted by CyTOF) and the occurrence of severe irAEs ( Fig.7D and E). CD4 T cell clusters 5 and 3 became the highest ranked pair. Using the above statistical method, the empirical P value of this result was calculated to be 0.002. Figure 3C Seurat FindMarkers was applied with default parameters to the CD4 T 5+3 population versus other CD4 T cell states to identify differentially expressed genes (DEGs).

[0190] To assess the relative usefulness of unsupervised clustering in describing the cellular determinants of irAE development, our scRNA-seq dataset was mapped to a PBMC atlas of 161,764 cells across 6 major lineages and 27 fine-grained subgroups defined by scRNA-seq and co-detected for more than 220 protein markers using a reference-guided annotation framework in Seurat v.4.0.1 (Azimuth). First, the query dataset was preprocessed following the quality control steps described above, yielding 24,807 cells. The query dataset was then normalized by SCTransform, FindTransferAnchors was applied to the query and reference datasets using a precomputed supervised PCA transformation with 50 dimensions, and MapQuery was then applied to map cell type labels and UMAP structures from the reference dataset to the query dataset.

[0191] Among the 27 cell states identified by Azimuth ( Fig. 8A ), CD4 TEM was most strongly associated with the occurrence of severe irAEs and was most correlated with CD4 TEM cells counted by CyTOF ( Figure 8C Of the two other CD4 TEM-like subsets identified by Azimuth (CD4 CTL, proliferative CD4), proliferative CD4 showed the highest expression of HLA-DX and the lowest expression of SELL ( Fig.8D ), which is consistent with an activated CD4 TEM phenotype. Furthermore, when Azimuth-derived protein expression was examined from antibody-derived labeling data, only CD4 TEM and CD4 proliferative states showed markers of TEM cells (CD45ROhiCD45RAloCD27lo; Fig. 8E In fact, the combined CD4 TEM and proliferative CD4 population was most associated with severe irAEs ( Figure 8C A hypergeometric test was applied to assess the overlap in cellular barcodes between the combined CD4 TEM+CD4 proliferative population (Azimuth) and states defined by de novo clustering. CD4 T 5+3 emerged as the top hit (Benjamini-Hochberg adjusted P = 2.5 × 10 -7 ). Despite strong overlap between unsupervised and supervised methods, CD4 T 5+3 was more strongly associated with severe irAE development and CyTOF than the cluster labeled by reference-guided annotation ( Fig.8F ).

[0192] Bulk RNA-seq library preparation, sequencing, and quantification.

[0193] The cryopreserved cell suspension was thawed as described above. RNA was subsequently extracted using the RNeasy PowerLyzer Tissue and Cell Kit (QIAGEN) and the quality was assessed using the 2100 Bioanalyzer System (Agilent Technologies). All samples were of sufficiently high quality for TruSeq RNA Exome Analysis (DV200>30%) and were prepared using the TruSeq RNA Exome Kit (Illumina) according to the manufacturer's instructions. After hybrid capture, cDNA libraries were pooled and sequenced on a HiSeq 2500 instrument (Illumina) using 2×150 bp double-end reads, with a target of 20 million to 25 million reads per sample. Raw reads were quantified using Salmon v.0.12.0 using the GENCODE v.29 reference transcriptome; the following command line parameters were used with other default parameters: --seqBias --gcBias --posBias --validateMappings --rangeFactorizationBins 4. Read counts were normalized to transcripts per million (TPM) using tximport v.1.10.1. Only samples with a mapping rate ≥ 60% and successful TCR assembly (see V(D)J receptor analysis and clonotype analysis below) were included for further analysis, with the exception of 3 samples with a mapping rate > 40% (but < 60%) and successful TCR assembly (included). In batch cohorts 1 and 2, a total of 53 sequenced samples (88%) met these criteria ( Figure 1 ).

[0194] Batch RNA-seq deconvolution.

[0195] To determine the leukocyte composition in bulk RNA-seq analysis of PBMCs, CIBERSORTx v.1.0.41 (https: / / cibersortx.stanford.edu) with the LM22 signature matrix was applied to the TPM matrix of each cohort ( Figure 1 CIBERSORTx applied B-mode batch correction separately and did not perform quantile normalization for each sequencing batch. LM22 (composed of highly optimized reference profiles for distinguishing 22 functionally defined human hematopoietic subsets) has been extensively validated for flow cytometry for accurate enumeration of leukocyte subsets in whole blood and PBMCs, whether analyzed by RNA-seq or microarray. Fig.11AThe performance of CIBERSORTx and LM22 activated CD4 TM cell analysis was further confirmed in this work by comparisons between CIBERSORTx, mass cytometry, flow cytometry, and scRNA-seq using PBMC samples from patients with melanoma ( Fig. 11B , 11C , 11D, 11E, and 11F). All LM22 subsets except granulocyte and macrophage subsets were evaluated in this work (n = 15; Figure 4A ), and the relative scores for each sample were renormalized to sum to 1. Although a total of 15 subgroups were assessed, 2 subgroups (regulatory T (Treg) cells, γδ T cells) were sparsely detected by CIBERSORTx and could not be identified by Figure 4A The Wilcoxon rank sum test was used to evaluate the

[0196] V(D)J receptor analysis and clonotype analysis.

[0197] For the single-cell discovery cohort, raw scV(D)J-seq reads were mapped to the reference refdata-cellranger-vdj-GRCh38-altsensembl-4.0.0 using Cell Ranger v.3.1.0, and the resulting clonal type assemblies were downloaded from Loupe v(D)J Browser v.3.0.0 (10x Genomics). Given that activated TM cells are derived from clonal expansions, the former are expected to have lower TCR diversity than their naive counterparts, provided that (1) cells from both populations are sampled equally (i.e., their counts are equal) or (2) changes in total T cell counts are normalized ( Fig. 10A However, by ignoring changes in total T cell frequencies, this sampling ignores richness (the number of unique species (clonotypes) in a population and a key factor underlying immune repertoire diversity). Therefore, the Shannon entropy, an information-theoretic metric that combines evenness and richness in a single measure, is primarily used to characterize immune repertoire diversity in this work. Fig. 10A ). For each evaluable patient sample in the single-cell discovery cohort ( Figure 1 and 2A), randomly sampling the TCR clonal repertoire (without replacement) to balance the number of evaluable PBMC cells across patients while accounting for technical differences in TCR recovery. To maximize the TCR clonal repertoire available for sampling, patients with <100 TCR clones were excluded (n=4; YUTAUR, YUTORY, YUHERN, and YUTHEA). We then calculated the Shannon entropy (Rpackage vegan v.2.5-6 (ref.78)) relative to total PBMCs for each T cell subset and averaged the values ​​obtained for the remaining 9 patients across 100 iterations of this procedure ( Figure 3E , 10B and 10D-F). Shannon entropy analysis of scBCR clonotypes on the IGK, IGL, and IGH chains of the same nine patients was performed as described above ( Figure 10G ).

[0198] For batch cohorts 1 and 2, after aptamer sequence trimming using Skewer v.0.2.2, TCR clonotypes were assembled and quantified with MiXCR v.3.0.125 using the following command: mixcr align -p rna-seq -s hsa -O allowPartialAlignments=true data_R1.fastq.gz data_R2.fastq.gzalignments.vdjca. For each patient sample, TCR clonotype diversity was measured overall for TCR-α and TCR-β chains using Shannon entropy (R package vegan v.2.5-6) ​​and compared between patients based on irAE severity ( Figure 4B , 4C , 12A, and 12C). The Gini-Simpson index, calculated using the R package immunarch v.0.6.5 (https: / / doi.org / 10.5281 / zenodo.3367200), was also applied to assess batch TCR diversity according to irAE severity ( Fig. 12B and 12D ). It is worth noting that TCR richness is a key component in calculating Shannon entropy and Gini-Simpson index.

[0199] Kinetic analysis of T cell clones from bulk PBMCs.

[0200] Bulk TCR-β chain analysis was performed on paired pre-treatment and early on-treatment PBMCs from 15 patients treated with combination ICIs. No patients underwent on-treatment peripheral blood collection after the onset of severe irAEs. Genomic DNA was extracted using the DNeasy Blood and Tissue Kit (QIAGEN) and submitted to survey-resolution immunoSEQ (survey-resolution immunoSEQ, Adaptive Biotechnologies). Data from productive TCR-β chain rearrangements were exported using the immunoSEQ Analyzer online tool, and TCR-β repertoire richness and diversity were assessed using Pielou uniformity, with increased 1-uniformity and increased clonality associated. Pielou uniformity results from immunoSEQ analysis were compared with bulk RNA-seq (MiXCR), which showed consistency ( Fig.15A We also verified that all pre- and on-treatment samples were correctly paired by cross-comparison of TCR-β CDR3 sequences. Clonal expansion was inferred by analyzing the difference in clonality between paired on- and pre-treatment time points (defined as 1-Pielou uniformity in each sample). Figure 5B and 15B More specifically, to calculate the change in clonality from baseline, pre-treatment clonality was subtracted from on-treatment clonality in a pairwise manner, thereby normalizing all pre-treatment samples to zero ( Figure 5B , left). We also analyzed Figure 5B data without targeting samples under treatment Fig. 15B Pairwise pre-treatment sample normalization in .

[0201] To assess the absence of severe irAEs, the extent of clonal expansion (expressed as δ) was divided into tertiles using the R package dplyr v.1.0.7 ( Figure 5D and 15H ). This yielded the following groups: no clonal expansion, δ < 0, n = 5; moderate, 0 < δ < 0.009, n = 5; and high clonal expansion, δ > 0.009, n = 5. These thresholds were applied to the full immunoSEQ cohort (n = 15; Figure 5D ) and patients in whom blood samples were obtained on day 1 and <1 month after ICI treatment (n=7; Fig.15H ). Furthermore, when represented in rank space, the extent of clonal expansion and the time to development of severe irAEs were significantly correlated in a Cox regression model and were independent of the time between blood draws, the number of productive TCR clones detected, and the age and sex of each patient.

[0202] Analysis of persisting T cell clones.

[0203] Paired pre-treatment peripheral blood scRNA-seq and scTCR-seq were performed on three patients (YUALOE, YUNANCY, and YUHONEY) Figure 5B ), these patients experienced severe irAEs and had varying levels of clonal expansion ( Figure 5C and 15D , 15E, 15F, and 15G). Of note, samples from these three patients had not been previously analyzed by scRNA-seq or scV(D)J-seq in the single-cell discovery cohort. Sequencing libraries were generated and processed for quality control identically to those described for the single-cell discovery cohort. Mapping was performed using Cell Ranger v.5.0.1.

[0204] To analyze persisting clones (defined as productive TCR-CDR3 nucleotide sequences shared between paired pre-treatment and on-treatment blood samples), the immunoSEQ data were interrogated for shared clonotypes where there were at least 2 templates in one blood draw (pre-treatment or on-treatment) and at least 1 template in the other (accounting for an average of 60% of all shared clones). This allowed one to preferentially focus on expanding or contracting persisting clones. The resulting sequences were cross-referenced with TCR-β CDR3 nucleotide sequences from the pre-treatment scTCR-seq library, which were further cross-referenced with scRNA-seq data and filtered for cells annotated as T cells by Azimuth (applied as described above) ( Fig.15D ). A total of 1504 single-cell transcriptomes with paired immunoSEQ clonotype data were identified. A significant Spearman correlation was observed between pre-treatment single-cell and immunoSEQ TCR clonotype frequencies for each patient (ρ>0.59; P0 and CD8A / B=0 for CD4 T cells; CD8A or CD8B>0 and CD4=0 for CD8 T cells). Overall, this approach could unambiguously label 69% of all cross-referenced clonotypes ( Fig.15E ).for Figure 5C The figure shown calculates the average log2 fold change between CD4 T 5 and 3 compared to the remaining CD4 T cell clusters in the single cell discovery cohort ( Figure 7B ), and the top 20 genes were selected for subsequent analysis. The enrichment of this gene set was determined using single-sample GSEA (R package escape v.1.0.1), which was applied to T cells labeled by Azimuth or T cells labeled as persistent CD4 / CD8 T cells as described above. Fig.15F The analysis shown in Figure 4 and G shows the frequency of persistent T cell clones generated by immunoSEQ grouped into CD4 and CD8 T cells, and the frequency of each clonotype ( Figure 15G ) or in general ( Fig.15F ) showed differences in generation frequency and compared with bulk clonal expansion from baseline ( Figure 5B ).

[0205] A comprehensive model for predicting the occurrence of irAEs.

[0206] The abundance of activated CD4 TM cells and the diversity of bulk TCR clonotypes were associated with the occurrence of severe irAEs ( Figure 4A and B). Therefore, integrative modeling was explored as a means to improve performance. Although several techniques were evaluated, including nonlinear modeling with random forests, logistic regression (glm in R) achieved comparable performance and was chosen due to the relative simplicity and robustness of generalized linear models. Prior to training, each feature was tested for outliers in batch cohorts 1 and 2 using the ROUT test with a false discovery rate = 10%. Among the 88 data points (2 features × 53 samples), 3 outliers were detected, all from activated CD4 TM cells in batch cohort 1. Regardless of the training cohort, all detected outliers always came from these three samples. Therefore, for each integrative model, the maximum fraction, maxF, of activated CD4 TM cell levels was determined from all non-outlier samples in the training cohort. maxF was then used as an upper limit for all samples.

[0207] The composite model was trained in several ways to predict the occurrence of severe irAEs (grade 3+). These included training on bulk cohort 1 and testing on a held-out bulk cohort 2 ( Figure 4D , left); training for one therapy type and testing for another therapy type ( Figure 4D , right); and trained for batch cohorts using LOOCV. For all models evaluated by LOOCV, the analysis was repeated n times, where n is the total number of patients. In each iteration, the model was trained for each patient except the i-th patient, and the i-th patient left out was evaluated. To mitigate overfitting when LOOCV divided patients into high and low groups, we applied Youden's J statistic to determine the threshold that optimizes sensitivity and specificity in each training cohort, and then assigned the i-th patient left out based on this threshold.

[0208] Composite model scores were evaluated by receiver operating characteristic (ROC) analysis. Models trained to distinguish between those with and without severe irAEs were used to predict future occurrence of severe irAEs ( Figure 4D and 13A ), irAE level ( Figure 4C , 4E , 13B and 13E), the number of organ systems affected by irAEs ( Fig.13H-J) and the time to development of severe irAEs ( Figure 5A and Figure 14). They were evaluated in different patient subgroups ( Figure 4D and 13D ) and compared with pathways assessed in bulk RNA-seq and previously published biomarkers ( Fig. 13C The composite model was additionally validated at different irAE level thresholds ( Fig.13F ) and were tested separately by treatment type to predict the occurrence of irAEs ( Figure 4D , 5A , 13A, 13D-F, 8B, 8C).

[0209] Assessment of circulating leukocyte composition in autoimmune diseases.

[0210] Peripheral blood gene expression datasets analyzed by bulk RNA-seq or microarray (across 239 patients with SLE, 348 patients with IBD, and 191 paired healthy controls) were downloaded from the Gene Expression Omnibus (GEO). RNA-seq data from Hung et al. were downloaded as preprocessed expression matrices and TPM normalized prior to analysis. Affymetrix microarray datasets (n = 5) were downloaded as CEL files, MAS5 normalized (affy v.3.12 in R (ref. 82)), mapped to Entrez gene identifiers using custom chip definition files specific to each platform (http: / / brainarray.mbni.med.umich.edu / Brainarray / Database / CustomCDF / ) and converted to HUGO gene symbols. One dataset did not have the original CEL file available; instead, preprocessed expression data were obtained from GEO. In cases where multiple probe sets mapped to the same gene symbol, we selected the probe set with the highest mean log2 expression across samples for further analysis. In the dataset of Palmer et al., some samples identified as controls were from subjects with Escherichia coli infection, celiac disease, or progression to Crohn's disease; these samples were excluded from the analysis. For replicates in the dataset of Carpinto et al., the most recent sample was selected. For the dataset of Peters et al., only pretreatment (week 0) blood samples from patients with Crohn's disease were further analyzed. CIBERSORTx49 was applied with LM22 to the bulk RNA-seq dataset of Hung et al. as described above, while the microarray datasets were either run with (1) quantile normalization and B-mode batch correction (non-HG-U133 platform) or (2) quantile normalization and no batch correction (HG-U133 platform). Leukocyte subsets were limited to monocyte subsets found in peripheral blood (granulocytes and macrophages were omitted) and were renormalized to sum to 1 for each sample.

[0211] In each dataset, a two-sided unpaired Wilcoxon rank sum test was applied to assess the levels of each leukocyte subset in peripheral blood between individuals with the disease and healthy controls from the same study ( Fig.16 ). The resulting P values ​​were converted to two-sided z scores while taking into account the directionality of the association. Within a given disease phenotype (SLE or IBD), z scores were combined across datasets using Lipták’s method weighted by sample size ( Fig.16 Finally, the SLE- and IBD-specific meta z scores were combined by Stouffer's method ( Fig.16) to generate a pan SLE / IBD meta z-score for each leukocyte subset ( Figure 6 ).

[0212] Candidate toxicity biomarkers from previous literature and pathway analysis.

[0213] The composite model was benchmarked against previously published irAE biomarkers and enriched pathways for severe irAE prediction ( Fig. 13C ). Each candidate biomarker was evaluated separately in bulk cohorts 1 and 2 by determining the AUC by ROC analysis. In this study, the following pre-treatment irAE biomarkers measured by protein expression in the literature were evaluated by RNA surrogates in peripheral blood: ADPGK and LCP1, which we evaluated individually and with bivariate linear regression; CD74 and GNAL15 expression; and CYTOX score, which was evaluated as the geometric mean expression of genes encoding the same 11 cytokines (CSF3, CSF2, CX3CL1, FGF2, IFNA2, IL12A, IL1A, IL1B, IL1RA, IL2, IL13). In addition, pre-ranked GSEA v.4.1.0 by GSEAPreranked v.7.1.0 was applied to identify the most irAE-enriched pathways in bulk cohorts 1 and 2 from the molecular signature database v7.4hallmark pathway set. As input, transcriptome-wide gene lists were defined for bulk cohorts 1 and 2, which were ranked by log2 fold change between patients who developed severe irAEs versus those who did not. Gene sets with q < 0.25 were considered statistically significant. The two most enriched gene sets (MYC_TARGETS_V1; OXIDATIVE_PHOSPHORYLATION) for patients with severe irAEs versus those without were compared to the composite model in bulk cohorts 1 and 2 ( Fig. 13C ).

[0214] statistics.

[0215] All statistical tests were two-sided unless otherwise stated. The Wilcoxon rank-sum test was used to assess statistical differences between the two groups. When >2 groups were assessed simultaneously, the nonparametric Kruskal-Wallis test was used. The Benjamini-Hochberg method was applied for multiple hypothesis testing unless otherwise stated. As described above, a permutation scheme was implemented to assess the association of scRNA-seq clustering with the occurrence of severe irAEs and CyTOF CD4 TEM abundance. Fisher's exact test was applied to assess statistical differences between two categorical variables. ROC analysis was performed to assess classification accuracy, which was quantified by AUC. The statistical significance of AUC was determined by a two-sided z test. After ROC analysis, Youden's J statistic was used to identify the optimal cut point. Linear consistency was determined by Pearson (r) or Spearman (ρ) correlation, and a two-sided t-test was used to assess whether the results were significantly nonzero. Kaplan–Meier and Cox regression analyses were used to assess covariates for the time to develop severe irAEs. The significance level and HR for Kaplan–Meier analysis were determined using a two-sided log-rank test. Figure 5A The composite model and related analyses in Figure 14 included patients from bulk cohorts 1 and 2 ( Figure 1 ), with the exception of two patients (YUDIME and YUMEDIC) who did not develop severe irAEs but experienced early disease progression that led to a switch in therapy before three months had elapsed. These two patients were included in the other analyses because they each received immune checkpoint blockade for 63 days (2.1 months), during which time 76% of all severe irAEs in the patient population occurred.

[0216] For Cox regression, results were analyzed based on the Wald statistic (z score), and significance was assessed by the Wald test. The proportional hazards assumption for each covariate included in the Cox regression was confirmed by assessing the Schoenfeld residuals before analysis. Comprehensive statistical analyses were performed using the method of Lipták and Stouffer when appropriate. Sample size calculations for bulk cohorts 1 and 2 were performed using pwr v.1.3-0 in R86. In the single-cell discovery cohort, the effect size for the association between CD4 TEM cell abundance (CyTOF) and the development of severe irAEs was 1.99 ( Figure 2C and D). Batch cohorts 1 and 2 were intended to meet the effect size requirements of α = 0.05 and 1-β = 0.8, while emphasizing the specificity of batch cohort 1 (the number of patients without severe irAEs > the number of patients with severe irAEs) and the balance in batch cohort 2 (the number of patients without severe irAEs ≈ the number of patients with severe irAEs). All statistical analyses were performed using R v.3.5.1+ or Prism 8+ (GraphPad Software).

Claims

1. A method for predicting the likelihood of a severe immune-related adverse reaction (irAE) in a patient receiving immunotherapy, the method comprising: a. Obtaining a peripheral blood sample from the subject before receiving immunotherapy treatment; b. Quantify the abundance of activated CD4 memory T cells and the diversity of T cell receptors (TCR) in peripheral blood samples; and c. If the abundance of activated CD4 memory T cells combined with the amount of TCR exceeds a threshold, the patient is classified as likely to develop severe irAR.

2. The method of claim 1, further comprising determining the threshold by reference to known clinical standards.

3. The method of claim 1, wherein the abundance of the activated CD4 memory T cells and the diversity of the T cell receptor (TCR) are determined using at least one of: bulk RNA sequencing (CIBERSORTx and MiXCR), time-of-flight mass cytometry (CyTOF), TCR-β analysis, droplet-based scRNA-seq and scTCR-seq, and targeted RNA-seq using the RNAome targeting activated CD4 memory T cells.

4. A method for predicting the likelihood of a severe immune-related adverse reaction (irAE) in a patient receiving immunotherapy, the method comprising: a. obtaining a first peripheral blood sample from a subject before receiving immunotherapy treatment, and obtaining a second peripheral blood sample after administering immunotherapy to the patient; b. quantifying a first TCR diversity level from the first peripheral blood sample and a second TCR diversity level from the second peripheral blood sample; c. obtaining the extent of TCR expansion by subtracting the first TCR diversity level from the second TCR diversity level; d. If the extent of TCR expansion exceeds the threshold, the patient is classified as likely to develop severe irAR.

5. The method of claim 4, further comprising predicting the time of onset of severe irAR based on the extent of TCR expansion, wherein a higher degree of TCR expansion predicts an earlier onset of severe irAR.

6. The method of claim 4, wherein the first TCR diversity and the second TCR diversity are determined using at least one of: bulk RNA sequencing (CIBERSORTx and MiXCR), time-of-flight mass cytometry (CyTOF), TCR-β analysis, droplet-based scRNA-seq and scTCR-seq, and targeted RNA-seq using the RNAome targeting activated CD4 memory T cells.

Images