Markers for prediction of adverse outcomes of car-t therapy

Abstract

Adverse outcomes of CAR-T such as hematotoxicity with prolonged cytopenia and infectious complications represent a challenging clinical problem after CAR-T therapy; however, current predictive models rely on blood counts and general inflammatory lab markers (ferritin, CRP) only and lack mechanistic / functional insights. Prolonged cytopenias after CAR-T are associated with endothelial alteration, characterized by reduced ANG1, E-selectin and MMP-1, and increased ANG2:ANG1 ratio, VCAM-1 and Thrombomodulin early after CAR-T. It was found that Patients with high baseline sIL-2R and VCAM-1 not only show more prolonged neutropenia and a more aplastic neutrophil recovery but also experience more severe infectious complications. High baseline VCAM-1 is associated with significantly worse peak CD8 CAR-T cell expansion and separates MM patients with worse overall response (Figure 5, Figure s5) Baseline sIL-2R and VCAM-1 have high predictive value for adverse outcomes, such as prolonged neutropenia, severe infections and death, and can further improve existing models.

Description

[0001] Markers for prediction of adverse outcomes of CAR-T therapy

[0002] Field of invention

[0003] The present invention presents novel markers for prediction of adverse outcomes of CAR-T therapy.

[0004] Background of the invention

[0005] Chimeric antigen receptor (CAR)-T cell therapy has become a well-established immunotherapy for a variety of hematologic malignancies. However, the efficacy of this therapy even in patients with relapsed / refractory disease comes at the cost of adverse outcomes with unique toxicities. Besides cytokine release syndrome (CRS) and immune effector cell- associated neurotoxicity syndrome (ICANS), which received the most attention among adverse events during pivotal trials, immune effector cell-associated hematotoxicity (ICAHT) with prolonged cytopenias and infectious complications have emerged as even more relevant problems in clinical practice1. In fact, hematotoxicity has proven to be the most common highgrade toxicity after CAR-T cell therapy in the real-world setting2-4.

[0006] CAR-T cell associated hematotoxicity is most likely not the result of one single factor but more of an interplay of multiple variables such as an impaired hematopoietic stem cell (HSC) reserve, disruption of the bone marrow microenvironment by local and systemic inflammation, as well as clonal T cell expansion and T cell-B cell imbalances1. Prior myelotoxic treatments, natural ageing and interactions with the underlying disease can affect the HSC reserve and lead to deregulated hematopoiesis. Within the bone marrow HSCs are predominantly located adjacent to sinusoids, where a special niche (microenvironment) primarily composed of endothelial cells, perivascular stromal cells and mesenchymal stromal cells promotes HSC maintenance by providing cytokines (e.g. SCF, CXCL12), forces and cell-mediated interactions5. CAR-T cells, when targeting BM infiltrating cells of the underlying disease or endogenous CD19+or BCMA+B-cell precursors within the BM, can induce cytokine release and local hyperinflammation, disrupting the delicate balance within the HSC niche. Furthermore, scRNAseq analysis of the bone marrow of patients with CAR-T associated hematotoxicity have revealed an association between clonally expanded CXCR1-high cytotoxic effector T cells with high expression of interferon-y (IFN-y) and development of prolonged cytopenias6. Besides that, increased systemic levels of inflammatory mediators such as IFN-y, tumor necrosis factor-a (TNF-a), interleukin-6 (IL-6), IL-18 and ferritin have been implicated with CAR-T associated hematotoxicity7’8. Strikingly, some of these systemic inflammatory mediators are already elevated before CAR-T infusion, highlighting the role of pre-existing inflammation in the development of CAR-T associated hematotoxicity4’8. Apart from a directly impairing effect on HSC, the signature of elevated mediators also indicates an overlap with immune effector cell- associated (IEC) haemophagocytic lymphophistiocytosis (HLH)-like syndrome (IEC-HS) as a potential pathomechanism of cytopenias9.

[0007] Patients with prolonged cytopenias more frequently require transfusions and medical consultations, which not only impair quality of life, but also significantly potentiate the risk of infectious complications in the case of severe and prolonged neutropenia and lymphopenia10. In line with this, a recent meta-analysis of over 7600 patients treated with CAR-T cells identified infections as the main driver of non-relapse mortality irrespective of the disease entity and CAR-T product11. Therefore, predictive models for adverse outcomes and thus proper risk assessment of patients are highly needed.

[0008] So far, several biomarkers and models have been proposed for prediction of adverse outcomes after CAR-T therapy4’12-18. The Endothelial Activation and Stress Index (EASIX) and several variations of it14’16’17’19were proposed for prediction of adverse outcomes after CAR-T by reflecting endothelial damage as a hallmark of CAR-T side effects1320. However, its main limitation is the use of non-specific surrogate blood markers that do not directly reflect endothelial damage. The CAR-Hematotox score was developed for prediction of prolonged neutropenia in lymphoma patients after CD19-directed CAR-T therapy (tisa-cel, axi-cel) and has subsequently been validated for CD19-directed CAR-T (brexu-cel) in mantle cell lymphoma and BCMA-directed CAR-T (ide-cel) in multiple myeloma (MM)4’21’22. At the cost of being an easy-to-use model, it only composes baseline blood counts (absolute neutrophil count, platelet count and hemoglobin) and inflammatory markers (C-reactive protein (CRP), ferritin) and therefore lacks mechanistic insights. Additionally, it did not prove its predictive capabilities in some studies, leaving room for improvement2324.

[0009] Our study aimed to characterize novel biomarkers to better predict adverse outcomes of CAR-T cell therapy and therefore improve risk stratification in clinical practice.

[0010] Summary of the invention

[0011] Adverse outcomes of CAR-T such as hematotoxicity with prolonged cytopenia and infectious complications represent a challenging clinical problem after CAR-T therapy; however, current predictive models rely on blood counts and general inflammatory lab markers (ferritin, CRP) only and lack mechanistic / functional insights.

[0012] Prolonged cytopenias after CAR-T are associated with endothelial alteration, characterized by reduced ANG1 , E-selectin and MMP-1 , and increased ANG2:ANG1 ratio, VCAM-1 and Thrombomodulin early after CAR-T.

[0013] It was found that Patients with high baseline slL-2R and VCAM-1 not only show more prolonged neutropenia and a more aplastic neutrophil recovery but also experience more severe infectious complications. High baseline VCAM-1 is associated with significantly worse peak CD8 CAR-T cell expansion and separates MM patients with worse overall response (Figure 5, Figure s5)

[0014] Baseline slL-2R and VCAM-1 have high predictive value for adverse outcomes, such as prolonged neutropenia, severe infections and death, and can further improve existing models.

[0015] BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1-7: Incidence and severity of hematologic toxicities after CAR-T. (Figs. 1-4) blood counts of neutrophils (Fig. 1), thrombocytes (Fig. 2), hemoglobin level (Fig. 3) and lymphocytes (4) before start of lymphodepletion and after CAR-T infusion shown as spaghetti plots depicting the individual course of every single patient (small lines) as well as the best-fit line from local polynomial regression (bold line) and its 95% confidence interval (dashed colored lines). The dashed grey lines show the border of respective grade 3 cytopenias. (Fig. 5) Mean cumulative duration of grade >3 neutropenia, thrombopenia, anemia and lymphopenia for MM and lymphoma patients. Whiskers indicate standard error of the mean (SEM). Statistical significance was determined by Mann Whitney U test (*p<0.05, **p<0.01 , ***p<0.001 , ****p<0.0001). (Fig. 6) Sankey diagram showing the time course from early ICAHT grade to late ICAHT grade for myeloma (up) and lymphoma (down) patients. NA = not available. (Fig. 7) Venn diagram showing the combined intersection (absolute numbers and percentage of the whole) of patients with >14 cumulative days of grade >3 anemia, thrombocytopenia, neutropenia and / or lymphopenia.

[0017] Figure 8-15: Post-infusion factors associated with severe hematotoxicity after CAR-T. (Fig. 8) Spearman correlation analysis of the influence of cytokines and flow cytometry markers early after CAR-T infusion (range day 3-7) and cumulative grade >3 anemia, grade >3 thrombopenia, grade >3 lymphopenia and grade >4 neutropenia. The Spearman correlation coefficient (r = scalebar) and the respective p values are depicted in the heatmap with red indicating positive correlation and blue indicating negative correlation (*p<0.05, **p<0.01 , ***p<0.001). (Figs. 9-15) Univariate analysis of the influence of ANG1 (Fig. 9), ANG2:ANG1 ratio (Fig. 10) and MMP-1 (Fig. 11) on prolonged thrombopenia, VCAM-1 (Fig. 12), ANG1 (Fig. 13) and Thrombomodulin (Fig. 14) on prolonged anemia as well as E-Selectin (Fig. 15) on prolonged neutropenia, after CAR-T therapy. Best-fit line (bold line) and 95% confidence intervals (dashed line) were calculated by simple linear regression. The Spearman correlation coefficient (r) and respective p- and q-values (after adjustment for multiple testing) are depicted for each marker. (B-H) shows all markers that fulfilled a q-value cutoff of <0.1.

[0018] Figure 16-24: Baseline factors associated with severe hematotoxicity after CAR-T. (Fig. 16) Spearman correlation analysis of the influence of cytokines and flow cytometry markers at baseline (before start of lymphodepletion) and cumulative grade >3 anemia, grade >3 thrombopenia, grade >3 lymphopenia and grade >4 neutropenia. The Spearman correlation coefficient (r = scalebar) and the respective p values are depicted in the heatmap with red indicating positive correlation and blue indicating negative correlation (*p<0.05, **p<0.01 , ***p<0.001). (Fig. 17-24) Univariate analysis of the influence of baseline VCAM-1 (Fig. 17), soluble IL-2 receptor (Fig. 18), CRP (Fig. 19), Ferritin (Fig. 20) and baseline tumor burden (Fig. 21) on prolonged neutropenia and baseline ANG1 (Fig. 22) as well as baseline VCAM-1 (Fig. 23) and the percentage of intermediate monocytes (Fig. 24) on prolonged thrombopenia and anemia. The influence of baseline tumor burden was analyzed by simple logistic regression and quantified by the likelihood ratio test (G2). For the other variables best-fit line (bold line) and 95% confidence intervals (dashed line) were calculated by simple linear regression. The Spearman correlation coefficient (r) and respective p- and q-values (after adjustment for multiple testing) are depicted for each tested marker. (Figs. 17-24) shows all markers that fulfilled a q-value cutoff of <0.1.

[0019] Figure 25-30: Hyperinflammation and endothelial dysfunction prior to CAR-T is associated with worse outcome. Kaplan-Meier curves for overall survival (Figs. 25 / 27) and progression-free survival (Figs. 26 / 28) for patients with high and low slL-2R (Fig. 25, Fig. 26) and VCAM-1 (Fig. 27, Fig. 28) at baseline (before start of lymphodepletion). Patients with values >3 quartile of all patients were characterized as high (slL-2R > 4915.9 pg / ml; VCAM-1 > 2032.3 ng / ml). Forest plots showing the results of multivariate cox regression analysis on the influence of different covariates on OS (Fig. 29) and PFS (Fig. 30).

[0020] Figure 31 -35: High slL-2R and VCAM-1 prior to CAR-T is associated with more prolonged neutropenia, more aplastic neutrophil recovery and more severe infections. (Fig. 31) Mean cumulative duration of grade 4 neutropenia between patients with high and low slL-2R or VCAM-1 at baseline. Whiskers indicate standard error of the mean (SEM). Statistical significance was determined by Mann Whitney U test. (Fig. 32) Percentage of quick, intermittent and aplastic neutrophil recovery phenotypes between patients with high and low slL-2R or VCAM-1 at baseline. (Fig. 33) Distribution of infection grades between patients with high and low slL-2R or VCAM-1 at baseline. (Fig. 34) Overall response rates of patients with high and low slL-2R or VCAM-1 at baseline subdivided by percentage of patients achieving partial remission (PR) / very good partial remission (VGPR) and complete remission (CR). Statistical significance was determined by Fisher’s exact test (Figs. 32-34). (Fig. 35) Peak CD4 and CD8 CAR-T expansion (dO until d14) per pl peripheral blood between patients with high and low slL-2R or VCAM-1 at baseline as determined by flow cytometry. Statistical significance was determined by Mann Whitney U test. *p<0.05, **p<0.01 , ***p<0.001 , ****p<0.0001. Patients with values >3 quartile of all patients were characterized as high.

[0021] Figure 36-42: Predictive value of baseline slL-2R and VCAM-1 on adverse outcomes after CAR-T. Receiver operating characteristic (ROC) curves of the influence of baseline sIL- 2R (Fig. 36) and VCAM-1 (Fig. 37) on the occurrence of prolonged grade 4 neutropenia of >7 days. ROC curves of the influence of baseline slL-2R (Fig. 38) and VCAM-1 (Fig. 39) on the occurrence of severe infections (> grade 3). ROC curves of the influence of baseline slL-2R (Fig. 40) and VCAM-1 (Fig. 41) on the occurrence of death. For ROC curves the respective area under the curve (AUC) and p-value are depicted. The concentration of the respective marker shown in brackets is the cut-off value with maximized Youden Index. (Fig. 42) Tabular summary of logistic regression analysis as univariate regressions for the CAR-Hematotox score alone as well as adjusted logistic regressions for predictive models that include baseline slL-2R or VCAM-1 in addition to the CAR-Hematotox score. The estimate, p-value and Akaike information criterion (AIC) for each model are shown.

[0022] Figure 43-49: Incidence and severity of hematologic toxicities. (Fig. 43) Study workflow and sampling timepoints. (Fig. 44) Simple logistic regression analysis comparing cumulative duration of grade >4 neutropenia to occurrence of death. (Figs. 45-48) Simple logistic regression analysis comparing cumulative duration of grade >4 neutropenia (Fig. 45), grade >3 lymphopenia (Fig. 46), grade >3 anemia (Fig. 47) and grade >3 thrombopenia (Fig. 48) to occurrence of severe infections (> grade 3). The log-likelihood ratio (G2) and respective p-value are depicted. (Fig. 49) Mean cumulative duration of grade >3 neutropenia, thrombopenia, anemia and lymphopenia for responders (n=50) vs. non-responders (n=13). Whiskers indicate standard error of the mean (SEM). Statistical significance was determined by Mann Whitney U test (*p<0.05, **p<0.01, ***p<0.001 , ****p<0.0001).

[0023] Figure 50-53: Correlation analysis of cytopenias per disease. Spearman correlation analysis of the influence of cytokines and flow cytometry markers at baseline (before lymphodepletion) (Fig. 50-51) and early after CAR-T infusion (range day 3-7) (Figs 52-53) on cumulative grade >3 anemia, grade >3 thrombopenia, grade >3 lymphopenia and grade >4 neutropenia for MM (Figs. 50, 52) and DLBCL patients (Figs. 51 , 53). The Spearman correlation coefficient (r = scalebar) and the respective p values are depicted in the heatmap with red indicating positive correlation and blue indicating negative correlation (*p<0.05, **p<0.01 , ***p<0.001).

[0024] Figure 54-64: Association between patient characteristics, CRS and ICANS and prolonged neutropenia after CAR-T. Univariate analysis of the influence of age (Fig. 54), disease entity (Fig. 55), high-risk cytogenetics (MM patients only) (Fig. 56), international prognostic index (DLBCL patients only) (Fig. 57), sex (Fig. 58), prior lines of therapy (Fig. 59), prior autologous / allogeneic stem cell transplantation (Fig. 60), CRS grade (Fig. 61), ICANS grade (Fig. 62), tocilizumab (Fig. 63) and dexamethasone treatment (Fig. 64) on the occurrence of prolonged neutropenia after CAR-T. For continuous variables the spearman correlation coefficient (r), respective p-value as well as the best-fit line (bold line) and 95% confidence interval (dashed line) as determined by simple linear regression is depicted. Binary variables were analyzed by simple logistic regression and quantified by the likelihood ratio test (G2).

[0025] Figure 65-72: OS and PFS based on baseline SIL-2R and VCAM-1 for MM and DLBCL patients. Kaplan-Meier curves for overall survival (Figs. 65-67) and progression-free survival (Figs. 69-72) for MM (Figs. 65, 67, 69, 71) and DLBCL (Figs. 66, 68, 70, 72) patients with high and low slL-2R (Figs. 65, 68) and VCAM-1 (Figs. 66, 69) at baseline (before start of lymphodepletion). Patients with values >3 quartile of all patients were characterized as high.

[0026] Figure 73: Influence of high baseline slL-2R and VCAM-1 on Response in MM patients. Overall response rates of MM patients with high and low slL-2R or VCAM-1 at baseline subdivided by percentage of patients achieving partial remission (PR), very good partial remission (VGPR), complete remission (CR) and MRD negative complete remission (CRD MRD neg).

[0027] Figure 74-75: Influence of high baseline slL-2R and VCAM-1 on CRS and ICANS. Distribution of CRS (Fig. 74) and ICANS grade (Fig. 75) between patients with high and low slL-2R or VCAM-1 at baseline. Patients with values >3 quartile of all patients were characterized as high. Statistical significance was determined by Fisher’s exact test (*p<0.05).

[0028] Figure 76-84: Receiver operating characteristic curves analyzing the influence of predictive models on prolonged neutropenia, severe infection and death. Receiver operating characteristic (ROC) curves of the influence of baseline CAR-Hematotox score on the occurrence of prolonged grade 4 neutropenia of >7 days (Fig. 76), the occurrence of severe infections (> grade 3) (Fig. 77) and occurrence of death (Fig. 78). ROC curves of modified predictive models incorporating either baseline VCAM-1 (Figs. 79-81) or slL-2R (Figs. 82-84) on the occurrence of prolonged neutropenia, severe infections or death (n=55). For ROC curves the respective area under the curve (AUC) and p-value are depicted.

[0029] Figure 85-90: Association between CAR-Hematotox score features and prolonged neutropenia after CAR-T. Univariate analysis of the influence of CAR-Hematotox score (Fig. 85), baseline CRP (Fig. 86), baseline ferritin (Fig. 87), baseline hemoglobin (Fig. 88), baseline absolute neutrophil count (Fig. 89) and baseline platelet count (Fig. 90) on the occurrence of prolonged neutropenia after CAR-T (n=64). Univariate analysis of the influence of CAR- Hematotox score (n=38 MM, n=26 DLBCL), CAR-Hematotox score + baseline SIL-2R (n=31 MM, n=24 DLBCL) and CAR-Hematotox score + baseline VCAM-1 (n=31 MM, n=24 DLBCL) for MM (G-l) and DLBCL (J-L). The spearman correlation coefficient (r), respective p-value as well as the best-fit line (bold line) and 95% confidence interval (dashed line) as determined by simple linear regression is depicted.

[0030] Figure 91-102: Correlation and ROC Analysis of CAR-Hematotox alone and in addition with slL-2R and VCAM-1 for MM and DLBCL patients. Univariate analysis of the influence of CAR-Hematotox score (n=38 MM, n=26 DLBCL), CAR-Hematotox score + baseline SIL-2R (n=31 MM, n=24 DLBCL) and CAR-Hematotox score + baseline VCAM-1 (n=31 MM, n=24 DLBCL) for MM (Figs. 97-99) and DLBCL (Figs. 100-102). The spearman correlation coefficient (r), respective p-value as well as the best-fit line (bold line) and 95% confidence interval (dashed line) as determined by simple linear regression is depicted. (Figs. 97-102) Receiver operating characteristic (ROC) curves of the influence of baseline CAR-Hematotox score alone (Figs. 97, 100) and in addition with slL-2R (Figs. 98, 101) or VCAM-1 (Figs. 99, 102) on the occurrence of prolonged grade 4 neutropenia of >7 days for MM (Figs. 97-99) and DLBCL (Figs. 100-102) patients. For ROC curves the respective area under the curve (AUC) and p-value are depicted.

[0031] Figure 103-112: Association between baseline slL-2R and VCAM1 and other factors. Univariate analysis of the association between baseline slL-2R and baseline tumor burden (Fig. 103), baseline CRP levels (Fig. 104), as well as IL-1 p (Fig. 105),TNF-a (Fig. 106) and IFN-y (Fig. 107) after CAR-T. (Fig. 108) Receiver operating characteristic (ROC) curves of the influence of baseline slL-2R on a hyperinfl am matory response after CAR-T, characterized by either I L-1 p, TNF-a or IFN-y above the third quartile after CAR-T. Univariate analysis of the association between baseline VCAM-1 and baseline R-ISS (Fig. 109), baseline hemoglobin levels (Fig. 110), baseline D-Dimer (Fig. 111) as well as VCAM-1 (Fig. 112) after CAR-T. For correlation analysis the spearman correlation coefficient (r), respective p-value as well as the best-fit line (bold line) and 95% confidence interval (dashed line) as determined by simple linear regression is depicted. For ROC curves the respective area under the curve (AUC) and p-value are depicted.

[0032] Detailed description

[0033] The invention is set out in the appended set of the claims. The invention includes the following items.

[0034] [1] A cell comprising a nucleic acid encoding a chimeric antigen receptor (CAR) that expresses the CAR on its surface for use in a method of treating cancer in a subject, wherein, following the administration of the cell to the subject:

[0035] (a) the subject is determined to have

[0036] (i) reduced levels of ANG1 ;

[0037] (ii) an increased ANG2:ANG1 ratio;

[0038] (iii) reduced levels of MMP-1 ;

[0039] (iv) increased levels of thrombomodulin;

[0040] (v) increased levels of VCAM-1 ; and / or

[0041] (vi) reduced levels E-selectin; and (b) the treatment is modified to prevent or treat a hematotoxicity associated with the administration of the cell.

[0042] [2] A method of predicting or diagnosing hematotoxicity in a subject undergoing treatment with a cell comprising a nucleic acid encoding a CAR that expresses the CAR on its surface, the method comprising:

[0043] (a) determining that

[0044] (i) the levels of ANG1, MMP-1 and / or E-selection are reduced in a sample isolated from the subject following administration of the cell;

[0045] (ii) the levels of thrombomodulin and / or VCAM-1 are increased in a sample isolated from the subject following administration of the cell; and / or

[0046] (iii) the ratio of ANG2:ANG1 is increased in a sample isolated from the subject following administration of the cell; and

[0047] (b) classifying the subject as being at risk of or suffering from a hematotoxicity following the determination of (a).

[0048] [3] A method of treating cancer comprising:

[0049] (a) administering a therapeutically effective dose of cells comprising a nucleic acid encoding a CAR, wherein the cells express the CAR on their surface;

[0050] (b) predicting or diagnosing a hematotoxicity in accordance with the method of item [2]; and

[0051] (c) modifying the treatment to prevent or treat the hematotoxicity associated with the administration of the cell.

[0052] [4] The cell for use of item [1] or the method of item [3], wherein the treatment is modified by:

[0053] (i) reducing the dose of cells administered to the subject; and / or

[0054] (ii) administering a salvage therapy (e.g., a transfusion and / or administering G- CSF).

[0055] [5] A cell comprising a nucleic acid encoding a CAR that expresses the CAR on its surface for use in a method of treating cancer in a subject, wherein: (i) the VCAM-1 and / or slL-2R levels of the subject prior to treatment are below a threshold value;

[0056] (ii) the tumor burden of the subject prior to treatment is below a threshold value; and / or

[0057] (iii) the percentage of intermediate monocytes in the blood of the subject prior to treatment is below a threshold value; and / or

[0058] (iv) the ANG1 levels of the subject prior to treatment are above a threshold value.

[0059] [6] A method for assessing the likelihood that a subject would suffer a severe hematotoxicity if treated with a cell comprising a nucleic acid encoding a CAR that expresses the CAR on its surface, the method comprising:

[0060] (a) determining that i) the levels of V-CAM-1 and / or slL-2R in a sample isolated from the subject prior to treatment are below a threshold value;

[0061] (ii) the tumor burden of the subject prior to treatment is below a threshold value;

[0062] (iii) the percentage of intermediate monocytes in a blood sample isolated from the subject prior to treatment is below a threshold value; and / or

[0063] (iv) the levels of ANG1 in a sample isolated from the subject prior to treatment are above a threshold value; and

[0064] (b) classifying the subject as being at low risk of suffering from a severe hematotoxicity following the determination of (a).

[0065] [7] A method of treating cancer comprising:

[0066] (a) determining that the subject is at low risk of suffering from a severe hematotoxicity in accordance with the method of item [6]; and

[0067] (b) administering a therapeutically effective dose of cells comprising a nucleic acid encoding a CAR, wherein the cells express the CAR on their surface.

[0068] [8] The cell for use of item [5] or the method of item [6] or [7], wherein:

[0069] (i) the levels of V-CAM-1 are below the upper quartile of the patient population suffering from cancer (e.g., suffering the same cancer as the subject); (ii) the levels of slL-2R are below the upper quartile of the patient population suffering from cancer (e.g., suffering the same cancer as the subject);

[0070] (iii) the levels of V-CAM-1 are below about 2100 ng / ml (e.g., below 2032 ng / ml); and / or

[0071] (iv) the levels of slL-2R are below about 5000 pg / ml (e.g., below 4916 pg / ml).

[0072] [9] The cell for use of any one of items [1], [4], [5] and [8], or the method of any one of items [2]-[4] and [6]-[8], wherein the sample of the subject that is assessed is a blood or serum sample.

[0073]

[0010] The cell for use of any one of items [1], [4], [5], [8] and [9], or the method of any one of items [2]-[4] and [6]-[9], wherein

[0074] (i) the CAR targets a cancer antigen (e.g., the CAR is an anti-BCMA or anti-CD19 CAR);

[0075] (ii) the cell is an immune cell (e.g., a T cell); and / or

[0076] (ii) the subject suffers from a blood cancer (e.g., lymphoma or multiple myeloma).

[0077]

[0011] The cell for use of item

[0010] or the method of item

[0010] , wherein:

[0078] (i) the cell is idecabtagene vicleucel and / or the subject suffers from multiple myeloma; or

[0079] (ii) the cell is axicabtagene ciloleucel and / or the subject suffers from lymphoma.

[0080] In any embodiment of the present invention, the cell is ciltacabtagene autoleucel and / or the subject suffers from multiple myeloma.

[0081] In any embodiment of the present invention, the VCAM-1 and / or slL-2R levels are assessed in combination with an EASIX score, mEASIX score or CAR-Hematotox score.

[0082] Examples

[0083] Methods Study design and subjects

[0084] This study is part of the IMI2 / EU project imSAVAR. To this end, 64 consecutive patients treated between November 2021 and December 2023 with the most commonly used approved CAR- T products idecabtagene vicleucel (ide-cel (Abecma®), n = 38) or axicabtagene ciloleucel (axi- cel (Yescarta®), n = 26) were included in an observational study at Universitatsklinikum Wurzburg. Five patients with missing or incomplete data and three patients with early loss to follow up (two on day 7 and one on day 11 after CAR-T) were excluded from the analysis. Patient characteristics, specifically demographics, disease characteristics, pretreatments, laboratory and staging results, adverse events and survival outcomes were collected with approval from the institutional review board and informed consent from the patients.

[0085] Sampling and cell processing

[0086] Samples were collected at fixed time points, specifically before apheresis (BS1), before lymphodepletion (BS2), day 3 to 7 after CAR-T infusion (BS3) and day 14 after CAR-T infusion (BS4) (Figure s1). Peripheral blood and serum was collected in EDTA anti-coagulated and clotting activator containing blood collection systems (Sarstedt, Numbrecht, Germany), respectively, via venipuncture according to the standard of care. Serum was collected after clotting by centrifugation at 1000xg for 15 min at 8°C and stored at -80°C until further analysis. Peripheral blood mononuclear cells (PBMC) were isolated from EDTA-blood by density gradient centrifugation using Pancoll human separating solution (density: 1.077 g / ml; PAN- Biotech GmbH, Aidenbach, Germany). PBMC were counted using a Neubauer counting chamber and trypan blue for dead cell exclusion. PBMC were then cryopreserved for subsequent downstream analyses using a freezing protocol by 10X Genomics (Pleasanton, CA, USA). All procedures were performed in accordance with the Declaration of Helsinki and the national ethical standards.

[0087] Grading of adverse events

[0088] Treatment related adverse events (AEs) were classified according to the Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. CRS, ICANS and ICAHT were graded according to American Society for Transplantation and Cellular Therapy (ASTCT) and EHA / EBMT consensus grading, respectively2526. Early and late ICAHT was classified by occurrence within the first 30 days or after day +30 after CAR-T infusion, respectively. Neutrophil recovery phenotypes were classified as quick, intermittent and aplastic according to Rejeski et al.4. For calculation of cumulative cytopenia days, the intervening days between assessments were only counted if both values were below the respective threshold. CRS, ICANS, cytopenias and infectious complications were only analyzed from day of CAR-T infusion until disease progression or start of new treatment regimens. Infectious complications were defined as bacterial, viral or fungal based on microbiology / virology test results, histopathologic data or as clinical syndrome of infection ± supportive imaging results (e.g. pneumonia). Fever alone after CAR-T cell administration without evidence of clinical symptoms or microbiological or imaging evidence was not considered an infection.

[0089] Flow cytometry

[0090] Peripheral blood was freshly analyzed at the respective time points by multi-color flow cytometry including the following markers: CD45, CD3, CD4, CD8, CD19, CD16, CD56, HLA- DR, CD14, CD11 b, CD11c, CD33 and reagents to detect CD19- or BCMA-CAR-positive cells, respectively (Table s4). Surface markers were stained directly in whole blood samples for 30 minutes (min) at room temperature (RT). After one washing step with MACSQuant® Running Buffer (Miltenyi Biotec, Bergisch Gladbach, Germany), red blood cells were lysed by incubation in 1X BD Pharm Lyse™ Lysing Buffer (BD Biosciences, Heidelberg, Germany) for 15 min at RT as recommended by the manufacturer. After an additional washing step, live / dead cell staining was carried out using the 7-AAD staining solution from Miltenyi Biotec. For CAR-T cell detection, whole blood was incubated with 1X BD Pharm Lyse™ Lysing Buffer for 15 min at RT to lyse red blood cells and washed once with MACSQuant® Running Buffer. Then, samples were incubated with the respective CAR detection reagent for 10 minutes at RT. After that, samples were washed and stained with antibodies for the following surface markers: CD45, CD3, CD4 and CD8 for 10 minutes at RT. After a final washing step, live / dead cell staining was carried out with the 7-AAD staining solution from Miltenyi Biotec. Cells with the forward and side light scatter properties of leukocytes were acquired on a MACSQuantIO analyzer. Data was analyzed gated on CD45+, live single cells with FlowJo (version 10.8.1) (Figure s11 ). Immune cell subsets were analyzed as frequencies and absolute cell counts of CD45+, live leukocytes.

[0091] Cytokine profiling

[0092] Cytokine profiling of 46 different analytes (complete list is shown in Table s2) was performed on cryopreserved serum samples of 55 patients (MM n = 31 ; DLBCL n = 24) at the time point before lymphodepletion (BS2) and on day 3 to 7 after CAR-T infusion (BS3). Out of the 64 patient cohort, only patients with incomplete sampling for the respective time points (6 patients) and a loss to follow-up before day 80 (3 patients) were excluded. Multiplex assays were purchased from ThermoFisher Scientific (Waltham, USA) and BioTechne (Minneapolis, USA) and were carried out according to the manufacturer’s protocol. Assays were analysed using a MAGPIX system and ProcartaPlex Analysis Application (Version 3.1 .1). Statistical analysis

[0093] Association between continuous variables was analyzed by calculating the Spearman correlation coefficient (r). Binary variables were analyzed by simple logistic regressions and association was quantified by likelihood ratio test (G2). To adjust for multiple testing for multiplex cytokine and flow cytometry analysis q-values were calculated and a false discovery rate (FDR) cut-off <0.1 was set. Progression-free survival (PFS) and overall survival (OS) were calculated using the Kaplan-Meier method and compared by log-rank test. Multivariable analysis was performed using a Cox proportional hazard model to determine the influence of the disease entity, high-risk profiles (High risk cytogenetics (MM) or I PI >3 (DLBCL)), presence of Bulk or elevated LDH or development of an aplastic neutrophil recovery on OS and PFS. To determine statistical significance between groups nonparametric Mann-Whitney U test was used for comparison of absolute values and Fisher’s exact test for comparison of percentages. GraphPad Prism (version 9.4.1) or R statistical software (version 4.3.1) was used for statistical analysis.

[0094] Results

[0095] Cytopenias and infectious complications are the most common adverse events after CAR-T therapy

[0096] Sixty-four patients were included in the analysis of whom 38 patients had relapsed / refractory MM treated with ide-cel and 26 patients had relapsed / refractory B-cell lymphoma (DLBCL) treated with axi-cel (Table 1). Forty patients (63%) were male and the median age was 62 years (range 22-79 years). The MM patients included 22 (58%) patients with high-risk cytogenetics. MM patients received a median of five prior lines of therapy (range 2-10 prior therapies), including autologous and allogeneic stem cell transplantation (SCT) in 37 (97%) and 10 (26%) of the patients, respectively. Thirty-three MM patients (87%) were exposed to lenalidomide, pomalidomide, bortezomib, carfilzomib and daratumumab (penta-exposed). Within the DLBCL patients, eleven patients (42%) had an international prognostic index (I PI) >3. The DLBCL patients received a median of two prior therapy lines (range, 1-5 prior therapies), including five patients (19%) with prior autologous SCT. Fourteen DLBCL patients (54%) had primary refractory disease and twelve (46%) were refractory to second or subsequent therapy lines. Fifteen DLBCL patients (58%) showed a high baseline tumor burden before CAR-T therapy characterized by either elevated LDH levels or presence of bulky disease, compared to nine (24%) MM patients with elevated LDH levels at baseline. In summary, the characteristics of our cohort reflects a typical representation of patients currently treated with most commonly used approved CAR-T cell products for MM and DLBCL patients in the real-world setting.

[0097] Out of 64 patients, 61 patients (95%) developed any grade CRS, whereas high-grade CRS (grade >3) only occurred in 2 patients (3%) (Table 2). Eighteen patients (28%) developed any grade ICANS, of which seven patients (12%), exclusively within the DLBCL cohort, had grade >3 ICANS (6 grade 3, 1 grade 4 event). Compared to that, infectious complications and cytopenias occurred more frequently. Thirty-seven patients (58%) developed any grade infections and 16 patients (25%) had severe infectious complications (grade >3) after CAR-T therapy. The majority of patients developed a grade >3 neutropenia (97%) or lymphopenia (94%) and about half the patients developed a grade >3 anemia (50%) or thrombopenia (59%). In addition, 41 % developed prolonged grade >3 neutropenia and 16% developed prolonged grade >4 neutropenia of more than 14 days. Nineteen patients (30%) developed prolonged grade >3 thrombopenia of more than 14 days with a cumulative duration of grade >3 thrombopenia over two months (>68 days) in seven patients (11 %). The mean cumulative days of grade >3 thrombopenia, neutropenia and lymphopenia were 16.3 (range 0-131), 18.1 (range 0-73) and 43.9 (range 0-389). The relevance of cytopenias post CAR-T is further highlighted by the amount of patients requiring transfusions (50%) or G-CSF support due to low neutrophil count before (84%) or later than day 14 (36%) after CAR-T infusion (Table 2). Furthermore, prolonged grade >4 neutropenia post CAR-T was found to be significantly associated with death in logistic regression analysis (Fig. 44). In addition to that, prolonged grade >4 neutropenia, grade >3 lymphopenia and grade >3 anemia but not prolonged grade >3 thrombopenia were found to be significantly associated with severe infections after CAR-T therapy (Fig. 45-48).

[0098] To analyze cytopenias and hematopoietic recovery after CAR-T therapy, we closely monitored lab counts for every patient before and after CAR-T treatment (Fig.1-4). Although the best-fit line of all patients showed a biphasic curve for all cytopenias, several patients developed severe, prolonged but also late cytopenias.

[0099] Looking closer at the different disease entities, MM patients treated with ide-cel developed more prolonged thrombopenia (p = 0.034) but showed no significant differences in the duration of neutropenia, anemia and lymphopenia compared to DLBCL patients treated with axi-cel (Fig. 5). Additionally, whereas lymphoma patients treated with axi-cel and MM patients treated with ide-cel had comparable early hematotoxicity (early ICAHT >3 12% vs. 14%), MM patients tended to develop more severe late hematotoxicity (late ICAHT 13% vs. 4%, n.s. by Fisher’s exact test) (Fig. 6). Of the 7 MM patients not available (N / A) for late ICAHT classification, 3 had progressive disease and 3 were lost to follow-up with only one visit after d30 and one patient died of viral pneumonia (early ICAHT grade 2). Of the five DLBCL patients not available (N / A) for late ICAHT classification, four had progressive disease and one patient died of aortic valve endocarditis and pneumocystis jirovecii pneumonia (early ICAHT grade 3).

[0100] We next studied the overlap between prolonged (>14 days) grade >3 anemia, thrombopenia, neutropenia and lymphopenia (Fig. 7). Compared to prolonged anemia and thrombopenia, which rarely occurred alone, prolonged lymphopenia often occurred isolated. The greatest intersection was seen between patients developing both prolonged neutropenia and lymphopenia (15%) as well as between prolonged thrombopenia and neutropenia or lymphopenia (9% each). Finally, prolonged neutropenia in general showed the greatest intersection with other prolonged cytopenias (31 % with lymphopenia, 30% with thrombopenia, 16% with anemia). Of note, besides duration of grade >3 lymphopenia, there was no significant difference in the duration of other severe (grade >3) cytopenias between responders and nonresponders (Fig. 49). Severe hematotoxicity after CAR-T therapy is associated with endothelial alteration

[0101] To study factors associated with CAR-T associated hematotoxicity we in-depth characterized patients undergoing CAR-T therapy by sequential multi-color flow cytometry and multiplex cytokine analysis (Fig. 43). First, we analyzed the association between factors detected early after CAR-T infusion (day 3 to day 7) and the occurrence of prolonged cytopenias. Within this analysis, we found several correlating markers (Fig. 8), of whom six markers (E-Selectin, ANG1 , ANG2:ANG1 ratio, MMP-1 , Thrombomodulin and VCAM-1), which all are known to be associated with endothelial alteration, remained significant after adjustment for multiple testing (q-value <0.1) (Fig 9-15). Reduced ANG1 and concomitantly increased ANG2:ANG1 ratio exhibited a strong correlation especially with prolonged thrombopenia and anemia early after CAR-T (Fig. 9,10,13). Furthermore, other endothelial markers such as reduced levels of MMP- 1 and increased levels of Thrombomodulin and VCAM-1 were significantly associated with prolonged thrombopenia and anemia, respectively (Fig. 11,12,14). Although reduced ANG1 and increased VCAM-1 seemed to be also associated with prolonged neutropenia after CAR- T (Fig. 8), a reduced E-Selectin level was the only marker, which remained to be significantly linked to prolonged neutropenia after adjustment for multiple testing (Fig.15). Of note, whereas ANG1 and ANG2:1 ratio after CAR-T showed a strong and uniform correlation with prolonged anemia, thrombopenia and neutropenia in MM patients, in DLBCL patients these markers were only significantly correlating with prolonged thrombopenia and E-Selection was found to be the factor most strongly associated with prolonged neutropenia (Fig. 52,53).

[0102] After adjustment for multiple testing, none of the flow cytometry markers early after CAR-T remained significant. However, prolonged neutropenia showed a moderate correlation with increased NK cells (r = 0.31 , p = 0.015) and decreased dendritic cell (r = -0.35, p = 0.0069) percentages in the peripheral blood early after CAR-T infusion ((Fig. 8)

[0103] Endothelial alteration, proinflammatory status and high tumor burden at baseline are risk factors for prolonged neutropenia after CAR-T

[0104] Next, we investigated baseline factors associated with the occurrence of prolonged cytopenias after CAR-T treatment. Again, several markers showed a significant correlation (Fig. 16), of whom four markers, namely VCAM-1 , soluble IL-2-receptor (slL-2R), ANG1 and the percentage of intermediate monocytes, remained significant after adjustment for multiple testing (Fig. 17-24). Again, we could detect endothelial markers as significant baseline markers for prolonged cytopenias after CAR-T. For instance, high baseline VCAM-1 strongly correlated with prolonged neutropenia and was significantly associated with prolonged anemia (Fig. 17, 23). Additionally, low baseline ANG1 was linked to prolonged thrombopenia after CAR-T (Fig. 22).

[0105] Furthermore, elevated baseline slL-2R showed a strong correlation with prolonged neutropenia (Fig. 18). Together with other correlating inflammatory markers such as elevated baseline ferritin and CRP, which are currently included in the CAR-Hematotox score4, this highlights the prognostic significance of a proinflammatory status prior to CAR-T (Fig. 19, 20). Of note, also increased IL-6 levels before CAR-T infusion (day 0) moderately correlated with prolonged neutropenia (r = 0.36, p = 0.0045, n=62) (Data not shown).

[0106] We also analyzed the influence of baseline patient characteristics on prolonged cytopenias, and only found a high tumor burden, reflected by the presence of bulky disease or elevated LDH at baseline, to be significantly associated with prolonged neutropenia after CAR-T (Fig. 21). Strikingly, neither the disease entity, high-risk profiles, the patients’ sex, number of prior therapy lines or a prior autologous / allogeneic stem cell transplantation were linked with prolonged neutropenia (Fig. 55-60) The weak correlation of patient age and prolonged neutropenia (Fig. 54) did not persist when myeloma and lymphoma patients were considered separately. Besides an association between prolonged thrombopenia and presence of MM as the underlying disease (G2= 5.84, p = 0.016), no other relevant association was found between the mentioned baseline characteristics and prolonged anemia, thrombopenia or lymphopenia (Data not shown).

[0107] Since endothelial alteration has been reported in the context of severe CRS and ICANS1320, we analyzed their influence on severe hematotoxicity. Surprisingly we could not see any association between CRS / ICANS severity or duration nor their treatment (dexamethasone and tocilizumab repeats) and prolonged neutropenia, thrombopenia, lymphopenia or anemia (Fig. 61-64). However, like hematotoxicity, CRS and ICANS were associated with elevated baseline proinflammatory markers, including IL-1 p, IL-1a, TNF-a, and IFN-y, further supporting the role of a proinflammatory state as a shared pathophysiological feature across CAR-T-related toxi cities.

[0108] The percentage of intermediate monocytes at baseline, which positively correlated with prolonged anemia, was the only flow cytometry markers at baseline, who remained significant after adjustment for multiple testing (Fig. 24). However, a reduced percentage of CD8 T cells (r = -0.36, p = 0.0056), high CD4:CD8 ratio at baseline (r = 0.38, p = 0.0033, Data not shown) and high percentage of monocytes (r = 0.37, p = 0.005) could display potential markers for prolonged lymphopenia after CAR-T. Patients with high baseline levels of either slL-2R or VCAM-1 have worse overall survival after CAR-T

[0109] Since baseline slL-2R and VCAM-1 showed the strongest correlation with prolonged neutropenia, we further investigated the impact of these markers on patient outcome. We therefore separated our cohort in two groups of either high (>3Q) or low baseline (<3Q) values, with a threshold based on the third quartile of each marker (slL-2R > 4916 pg / ml; VCAM-1 > 2032 ng / ml). Strikingly, patients with high baseline levels of either slL-2R or VCAM-1 had significantly worse overall survival (Fig. 25, 26) after CAR-T therapy. In a multivariate cox proportional hazard regression model for overall survival only the occurrence of an aplastic neutrophil recovery phenotype but not the underlying disease entity, presence of high-risk profiles or high tumor burden had significant impact on patients’ survival (Fig. 29). In line with this, when MM and DLBCL patients were considered separately, the clear separation of patients with worse overall survival remained significant for VCAM-1 and, although it did not reach significance, also slL-2R still showed the same trend (Fig. 65-68. Compared to the impaired overall survival, neither of the markers seem to affect progression-free survival after CAR-T therapy (Fig. 26, 28). However, MM patients with high baseline VCAM-1 had significantly worse progression-free survival and MM patients with high baseline slL-2R showed a trend towards worse progression-free survival (p = 0.07) (Fig. 69, 71).

[0110] Patients with high baseline levels of either slL-2R or VCAM-1 have prolonged neutropenia, show a more aplastic neutrophil recovery and more severe infectious complications

[0111] We next sought to investigate the reasons for poorer outcomes of patients with high baseline slL-2R and VCAM-1. Comparing the two groups with high and low baseline levels of our markers we could see that, patients with high slL-2R and VCAM-1 had significantly longer duration of neutropenia, a higher percentage of aplastic neutrophil recovery and a higher percentage of severe infectious complications (grade >3) after CAR-T (Fig. 31-33). Compared to that, patients with high baseline levels of slL-2R and VCAM-1 did not show significant differences in response rates, although patients with high VCAM-1 levels showed a substantially reduced percentage of complete remissions and significantly lower peak expansion of CD8 CAR-T cells (Fig. 34, 35). Of note, when looking at MM patients only, patients with high baseline markers showed substantially reduced minimal residual disease (MRD) negative CR rates and patients with high baseline VCAM-1 had a significantly lower overall response (Fig. 73). No significant differences in the percentages of any grade or high-grade CRS or ICANS could be found between patients with high and low baseline slL-2R and VCAM-1 (Fig. 73, 74). slL-2R and VCAM-1 at baseline have predictive value for adverse outcomes of CAR-T and can further improve existing models

[0112] The primary aim of this study was to identify novel markers for adverse outcomes of CAR-T therapy and thus improve risk stratification of patients. We therefore tested the predictive capacity of baseline slL-2R and VCAM-1 on adverse outcomes of CAR-T therapy in receiver operating characteristic (ROC) analysis. Both markers showed predictive value for the adverse outcomes of prolonged neutropenia >7 days, the occurrence of severe infections (grade >3) and death (Fig. 36-41). Whereas baseline slL-2R proved to be a better predictor of prolonged neutropenias and severe infectious complications, baseline VCAM-1 proved to be the better predictor for death after CAR-T therapy. Finally, logistic regression analysis showed significantly improved predictive power of the CAR-Hematotox score with the addition of slL-2R for the prediction of prolonged neutropenia (>7 days) and severe infections (grade >3) and with the addition of VCAM-1 for the prediction of death (Fig. 42).

[0113] The integration of these biomarkers into existing risk models, including EASIX and mEASIX, also consistently improved predictive performance, as reflected by increased ALICs and significance levels across nearly all endpoints. When adjusting for EASIX or mEASIX, both biomarkers demonstrated improved prognostic value: slL-2R significantly strengthened associations with all endpoints, while VCAM-1 particularly reinforced the association with prolonged neutropenia. Taken together, these findings suggest that both biomarkers provide incremental prognostic information beyond existing risk scores.

[0114] Discussion

[0115] Our study highlights endothelial alteration and proinflammatory state as substantial risk factors for severe hematotoxicity and adverse outcomes after CAR-T therapy. We further identified baseline VCAM-1 and soluble IL-2R as valuable molecular markers reflecting these states and proved their applicability to predict adverse outcomes after CAR-T therapy.

[0116] Foremost, predictive markers need to be broadly applicable and reproducible. Therefore, we attempted to develop markers based on a heterogeneous cohort of MM and DLBCL patients, the tumor entities currently most commonly treated with CAR-T cells in adults11. Importantly VCAM-1 and slL-2R showed comparable trends for both MM and DLBCL patients (Fig. 50-53, Fig. 65-72, Fig. 91-102); however, they appeared to be even more pronounced markers in MM patients. To date the most established model to predict CAR-T associated hematotoxicity is the CAR-hematotox score, which initially was developed in DLBCL patients with subsequent validation in different tumor entities4’21’22 26. Although some of its single features did not show the same correlations in our cohort, the overall score significantly correlated with prolonged neutropenias and proved to be a valuable predictor (Fig. 76, Fig. 85-90). However, given its development in DLBCL patients only, we and others have observed that it tends to underperform in MM patients (Fig. 91-102)24. The addition of baseline VCAM-1 or slL-2R could not only compensate for this deficiency in MM patients, but also further improve the model’s predictive ability for DLBCL patients (Fig. 91-102) and thus improving its applicability for a broader spectrum of patients.

[0117] VCAM-1 is a cell adhesion molecule upregulated on blood vessels during endothelial activation27. By binding to its ligand VLA-4 (integrin a4pi) it mediates leukocyte adhesion and transendothelial migration. VCAM-1 can be proteolytically cleaved and released as soluble VCAM-1 (sVCAM-1) into circulation, a process that is upregulated by IL-1 p and TNF-a during inflammatory response28. Increased levels of sVCAM-1 have been associated with numerous pathological conditions such as cardiovascular disease, immunological disorders and cancer2729. More specifically, sVCAM-1 was found to be elevated in advanced stage nonHodgkin’s lymphoma and multiple myeloma where it has been associated with poorer survival3031. In contrast to these findings, in our cohort elevated sVCAM-1 levels only correlated with increased baseline R-ISS (Fig. 109), but did not show an association with Ann- Arbor stage, I PI , presence of Bulk or elevated LDH. Besides that, elevated levels of sVCAM-1 have also been observed in patients with myelodysplastic syndromes (MDS), where it is associated with disease severity32. In line with this, a defective vascular niche contributes to the pathogenesis of myelodysplasia3334. Together the prognostic impact of VCAM-1 could not only explained by providing insight into advanced disease stages but also by reflecting preexisting damage to the bone marrow niche.

[0118] The IL-2 receptor has a low-affinity dimeric variant consisting of IL-2RP (CD122) and IL-2Ry (CD132) and a high-affinity trimeric variant that also possesses IL-2Ra (CD25). High levels of the trimeric IL-2R are transiently expressed on activated T cells and constitutively expressed on regulatory T cells35. IL-2Ra can be shed and released as a soluble form (slL-2R), a process that has been described for activated T cells, regulatory T cells, dendritic cells and monocytes3536. Elevated levels of slL-2R have been observed in autoimmune and inflammatory diseases, infections, solid cancers and hematological malignancies and are thought to reflect sustained immune activation and a pro-inflammatory state35. In patients with untreated aggressive B-cell lymphoma, elevated slL-2R levels were associated with advanced disease stage (i.e. Ann Arbor lll / IV and IPI high-intermediate / high) and were linked to poorer overall survival3738. Compared to that newly diagnosed MM patients with high level of slL-2R had worse progression free survival but showed no differences in overall survival3940. These reports are in line with our findings that elevated baseline slL-2R was strongly associated with high tumor burden, reflected by either elevated LDH or bulky disease (Fig. 103). Furthermore, we could detect a strong correlation between baseline slL-2R and CRP levels, and high baseline slL-2R levels were significantly associated with elevated levels of pro-inflammatory cytokines IL-ip, TNF-a and IFN-y after CAR-T infusion (Fig. 104-107). All in all, elevated baseline slL-2R does not only characterize patients with high tumor burden and advanced disease stages but also reflects a pro-inflammatory baseline state and predicts a hyperinflammatory response after CAR-T (Fig. 108).

[0119] In summary, our study identified endothelial alteration and a pro-inflammatory state as hallmarks of severe hematotoxicity and adverse outcomes after CAR-T therapy. Furthermore, VCAM-1 and slL-2R are suitable markers to detect these unfavorable conditions and can improve currently existing predictive models.

[0120] Tables

[0121] Table 1 : Baseline patient characteristics

[0122] All patients Characteristics

[0123] Male - no. (%)

[0124] Female - no. (%)

[0125] Age - median (range)

[0126] Tumor entity - no. (%)

[0127] Multiple Myeloma

[0128] DLBCL

[0129] Disease stage / risk profile

[0130] MM

[0131] R2-ISS III - no.

[0132] High-risk cytogenetics (del17p, t4; 14, +1q) - no. (%) 22 (58%)

[0133] Extramedullary disease - no. (%)

[0134] Elevated LDH - no. (%) 9 (24%) Ann Arbor > no. (%) 19 (73%)

[0135] IPI >3 - no.

[0136] Bulk or elevated LDH - no. (%) 15 (58%)

[0137] Prior therapies - MM Prior therapy lines - median (range)

[0138] Pomalidomide exposed - no. (%)

[0139] Carfilzomib exposed - no. (%)

[0140] Prior autologous SCT - no. (%)

[0141] Prior allogeneic SCT - no. (%)

[0142] Prior therapies - DLBCL Prior therapy lines - median (range)

[0143] Primary refractory disease - no. (%) 14 (54%)

[0144] Refractory to second or subsequent lines - no. (%)

[0145] Early relapse (<12 mo) after first line treatment - no. (%) 9 (35%)

[0146] Prior autologous SCT lymphoma - no. (%)

[0147] Products / infusion 64

[0148] Abecma (ide-cel) - no. (%)

[0149] Yescarta (axi-cel) - no. (%) 26 (41%) Table 2: Incidence and severity of adverse events after CAR-T.

[0150] Adverse events

[0151] Viral

[0152] Bacterial

[0153] Fungal

[0154] Unknown

[0155] Cytopenia any grade - no. (%)

[0156] Anemia CTC >3

[0157] Thrombopenia CTC >3 >3 for >14 days

[0158] Mean cumulative days of Thrombopenia <50 G / l (range) 16.3 (0-131) Transfusion required (pRBC or platelets) TPO agonists received 5 (8%) Lymphopenia >3 for >14 days 31 (48%)

[0159] Mean cumulative days of Lymphopenia <500 / pl (range) 43.9 (0-389) Neutropenia CTC >3 62 (97%) >3 for >14 days 26(41%)

[0160] Mean cumulative days of Neutropenia <1000 / pl (range) 18.1 (0-73)

[0161] Neutropenia

[0162] >4 for >14 days 10(16%)

[0163] Mean cumulative days of Neutropenia <500 / pl (range) 7.5 (0-50)

[0164] G-CSF Support d1-14

[0165] G-CSF Support >d14

[0166] Neutrophil Recovery type* quick

[0167] Intermittent 30 (47%) aplastic

[0168] Secondary malignancies

[0169] Follow-up Examples

[0170] In follow-up studies, the patient population was expanded to include a patient cohort of 23 subjects (n=23) treated with ciltacabtagene autoleucel (Cilta-cel, Carvykti®). Cilta-cel is a CAR T cell therapy wherein the CAR targets BCMA. The subjects had relapsed / refractory multiple myeloma.

[0171] Patient characteristics, bridging therapies13, laboratory and staging results, adverse events and survival outcomes were collected with approval from the institutional review board and informed consent from the patients. All patients received standard fludarabine / cyclophosphamide lymphodepletion. All procedures have been performed in accordance with the Declaration of Helsinki.

[0172] Samples were collected before apheresis (BS1), at baseline i.e. before lymphodepletion (BS2), early after CAR-T infusion (BS3, target day 9±1 for cilta-cel, to capture peak CAR-T expansion and inflammation based on kinetic profiles of the respective product), and on day 14 after CAR- T infusion (BS4). Peripheral blood was collected in EDTA anti-coagulated and serum in clotting activator containing blood collection systems (Sarstedt, Numbrecht, Germany). EDTA peripheral blood was processed freshly for flow cytometry. Serum was centrifuged at 1000xg for 15 min at 8°C and stored at -80°C until further analysis.

[0173] Adverse events were classified as described previously. Flow cytometry was performed in accordance with the methods described above.

[0174] Cytokine profiling of 42 markers was performed on cryopreserved serum samples of all patients at the time point before lymphodepletion (BS2) and early after CAR-T infusion (BS3). Analysis was performed employing three separate multiplex kits, analyzing 31 , 9 and 2 analytes per run, to ensure optimal assay performance and minimize cross-reactivity. Multiplex assays were purchased from ThermoFisher Scientific (Waltham, Massachusetts, USA) and BioTechne (Minneapolis, Minnesota, USA) and carried out according to the manufacturer’s protocol. Assays were analyzed using a Luminex MAGPIX® system (Austin, Texas, USA) and ProcartaPlex Analysis Application (Version 3.1.1). Out of range cytokine values were set to the highest or lowest standard value. Marker selection was based on literature research to reflect defined key events and risk factors of CAR-T associated hematotoxicity.

[0175] Statistical analysis was performed in accordance with the methods described above. It was also confirmed for cilta-cel early after CAR-T infusion (day 3-10, BS3) that reduced ANG1 and concomitantly increased ANG2:ANG1 ratio correlated with prolonged anemia, thrombocytopenia and neutropenia early after CAR-T. In addition, other endothelial markers such as reduced soluble P-selectin significantly correlated with prolonged thrombocytopenia and neutropenia, reduced VEGF-A with prolonged neutropenia and anemia, and increased VCAM-1 with prolonged thrombocytopenia and anemia. Finally, reduced soluble E-selectin levels showed the strongest but sole association with prolonged neutropenia and reduced MMP-1 was almost exclusively associated with prolonged thrombocytopenia after CAR-T.

[0176] It was also confirmed for cilta-cel that baseline VCAM-1 showed the strongest overall correlation with prolonged neutropenia (r = 0.46, p <0.0001) and anemia (r = 0.36, p = 0.0012) and baseline slL-2R showed a significant correlation with prolonged neutropenia.

[0177] Most importantly in follow-up analysis sVCAM-1 and slL-2R were identified as markers with independent additional prognostic values beyond existing scores such as EASIXand mEASIX. Incorporation of sVCAM-1 or slL-2R improved the prognostic value of EASIX, mEASIX and CAR-HEMATOTOX, as reflected by increased ALICs and significance levels in Reciever operating characteristic analysis for prolonged grade 4 neutropenia, severe infections and death after CAR-T therapy.

[0178] References Rejeski, ( K et al. Immune effector cell-associated haematotoxicity after CAR T-cell therapy: from mechanism to management. Review Lancet Haematol 11 , 459-70 (2024). Jain, T. et al. Hematopoietic recovery in patients receiving chimeric antigen receptor T-cell therapy for hematologic malignancies. Blood Adv 4, 3776-3787 (2020). Fried, S. et al. Early and late hematologic toxicity following CD19 CAR-T cells. Bone Marrow Transplantation 2019 54:1054, 1643-1650 (2019). Rejeski, K. et al. CAR-HEMATOTOX: a model for CAR T-cell-related hematologic toxicity in relapsed / refractory large B-cell lymphoma. Blood 138, 2499-2513 (2021). Ding, L., Saunders, T. L., Enikolopov, G. & Morrison, S. J. Endothelial and perivascular cells maintain haematopoietic stem cells. Nature 2012481:7382481 , 457-462 (2012). Strati, P. et al. Prolonged cytopenia following CD19 CAR T cell therapy is linked with bone marrow infiltration of clonally expanded IFNy-expressing CD8 T cells. Cell Rep Med 4, (2023). Rejeski, K. et al. Severe hematotoxicity after CD19 CAR-T therapy is associated with suppressive immune dysregulation and limited CAR-T expansion. Sci Adv 9, (2023). Juluri, K. R. et al. Severe cytokine release syndrome is associated with hematologic toxicity following CD19 CAR T-cell therapy. Blood Adv 6, 2055-2068 (2022). Nguyen, T. T. T., Kim, Y. T., Jeong, G. & Jin, M. Immunopathology of and potential therapeutics for secondary hemophagocytic lymphohistiocytosis / macrophage activation syndrome: a translational perspective. Experimental & Molecular Medicine 2024 56:356, 559-569 (2024). Rejeski, K. et al. Original research: The CAR-HEMATOTOX risk-stratifies patients for severe infections and disease progression after CD19 CAR-T in R / R LBCL. J Immunother Cancer 10, 4475 (2022). Cordas dos Santos, D. M. etal. A systematic review and meta-analysis of nonrelapse mortality after CAR T cell therapy. Nature Medicine 2024 1-12 (2024) doi : 10.1038 / s41591-024-03084- 6. Teachey, D. T. etal. Identification of predictive biomarkers for cytokine release syndrome after chimeric antigen receptor T-cell therapy for acute lymphoblastic leukemia. Cancer Discov 6, 664-679 (2016). Hay, K. A. et al. Kinetics and biomarkers of severe cytokine release syndrome after CD19 chimeric antigen receptor-modified T-cell therapy. Blood 130, 2295-2306 (2017). Greenbaum, II. et al. CRP and ferritin in addition to the EASIX score predict CAR-T-related toxicity. Blood Adv 5, 2799-2806 (2021). Shouse, G. et al. A validated composite comorbidity index predicts outcomes of CAR T-cell therapy in patients with diffuse large B-cell lymphoma. Blood Adv 7, 3516-3529 (2023). Pennisi, M. etal. Modified EASIX predicts severe cytokine release syndrome and neurotoxicity after chimeric antigen receptor T cells. Blood Adv 5, 3397-3406 (2021). Acosta-Medina, A. A. et al. Pre-lymphodepletion & infusion endothelial activation and stress index as predictors of clinical outcomes in CAR-T therapy for B-cell lymphoma. Blood Cancer Journal 2023 13:1 13, 1-3 (2023). Gagelmann, N. et al. Development and Validation of a Prediction Model of Outcome after 13- Cell Maturation Antigen-Directed Chimeric Antigen Receptor T-Cell Therapy in Relapsed / Refractory Multiple Myeloma. Journal of Clinical Oncology 42, 1665-1675 (2024). Greenbaum, II. et al. The Easix (Endothelial Activation and Stress Index) Score Predicts for CAR T Related Toxicity in Patients Receiving Axicabtagene Ciloleucel (axi-cel) for NonHodgkin Lymphoma (NHL). Blood 136, 17-18 (2020). Gust, J. et al. Endothelial Activation and Blood-Brain Barrier Disruption in Neurotoxicity after Adoptive Immunotherapy with CD19 CAR-T Cells. Cancer Discov 7, 1404-1419 (2017). Rejeski, K. etal. The CAR-HEMATOTOX score as a prognostic model of toxicity and response in patients receiving BCMA-directed CAR-T for relapsed / refractory multiple myeloma. J Hematol Oncol 16, 1-13 (2023). Rejeski, K. etal. The CAR-HEMATOTOX score identifies patients at high risk for hematological toxicity, infectious complications, and poor treatment outcomes following brexucabtagene autoleucel for relapsed or refractory MCL. Am J Hematol 98, 1699-1710 (2023). Corona, M. et al. Management of prolonged cytopenia following CAR T-cell therapy. Bone Marrow Transplantation 2022 57:1257 , 1839-1841 (2022). Oliver-Caldes, A. et al. CAR-T Related Cytopenia Profile in Relapsed / Refractory Multiple Myeloma: Results of Patients Treated with ARI0002h, an Academic BCMA-Directed CAR-T Cell. Blood 140, 10347-10349 (2022). Lee, D. W. et al. ASTCT Consensus Grading for Cytokine Release Syndrome and Neurologic Toxicity Associated with Immune Effector Cells. Biology of Blood and Marrow Transplantation 25, 625-638 (2019). Rejeski, K. et al. Immune effector cell-associated hematotoxicity: EHA / EBMT consensus grading and best practice recommendations. Blood 142, 865-877 (2023). Kong, D. H., Kim, Y. K., Kim, M. R., Jang, J. H. & Lee, S. Emerging Roles of Vascular Cell Adhesion Molecule-1 (VCAM-1) in Immunological Disorders and Cancer. Int J Mol Sci 19, (2018). Singh, R. J. R. et al. Cytokine stimulated vascular cell adhesion molecule-1 (VCAM-1) ectodomain release is regulated by TIMP-3. Cardiovasc Res 67, 39-49 (2005). Troncoso, M. F. etal. VCAM-1 as a predictor biomarker in cardiovascular disease. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1867, 166170 (2021). Christiansen, I., Sundstrdm, C., Kalkner, K. M., Bring, J. & Tdtterman, T. H. Serum levels of soluble vascular cell adhesion molecule-1 (sVCAM-1) are elevated in advanced stages of nonHodgkin’s lymphomas. Eur J Haematol 62, 202-209 (1999). Terpos, E. et al. Increased circulating VCAM-1 correlates with advanced disease and poor survival in patients with multiple myeloma: reduction by post-bortezomib and lenalidomide treatment. Blood Cancer Journal 2016 6:56, e428-e428 (2016). Passam, F. H. et al. Levels of soluble forms of ICAM and VCAM in patients with myelodysplastic syndromes and their prognostic significance. Clin Lab Haematol 26, 391-395 (2004). Teofili, L. et al. Endothelial Progenitor Cell Dysfunction in Myelodysplastic Syndromes: Possible Contribution of a Defective Vascular Niche to Myelodysplasia. Neoplasia 17, 401- 409 (2015). Mendez-Ferrer, S. etal. Bone marrow niches in haematological malignancies. Nature Reviews Cancer 202020:520, 285-298 (2020). Boyman, O. & Sprent, J. The role of interleukin-2 during homeostasis and activation of the immune system. Nature Reviews Immunology 2012 12:3 12, 180-190 (2012). Kniep, E. M., Strelow, I. & Lohmann-Matthes, M.-L. The monocyte interleukin-2 receptor light chain: production of cell-associated and soluble interleukin-2 receptor by monocytes. Immunology75, 299 (1992). Niitsu, N., lijima, K. & Chizuka, A. A high serum-soluble interleukin-2 receptor level is associated with a poor outcome of aggressive non-Hodgkin’s lymphoma. Eur J Haematol 66, 24-30 (2001). Goto, H. et al. Serum-soluble interleukin-2 receptor (slL-2R) level determines clinical outcome in patients with aggressive non-Hodgkin’s lymphoma: In combination with the International Prognostic Index. J Cancer Res Clin Oncol 131, 73-79 (2005). Lin, P. et al. Expression of SIL-2R in Patients with Multiple Myeloma and Its Clinical Significance. Journal of Experimental Hematology 30, 1797-1802 (2022). Wang, L. & Xia, Z. High Level of Soluble lnterleukin-2 Receptor in Serum Predicts Treatment Resistance and Poor Progression Free Survival in Multiple Myeloma. Blood 126, 5332 (2015).

Claims

CLAIMS1. A cell comprising a nucleic acid encoding a chimeric antigen receptor (CAR) that expresses the CAR on its surface for use in a method of treating cancer in a subject, wherein, following the administration of the cell to the subject:(a) the subject is determined to have(i) reduced levels of ANG1;(ii) an increased ANG2:ANG1 ratio;(iii) reduced levels of MMP-1;(iv) increased levels of thrombomodulin;(v) increased levels of VCAM-1; and / or(vi) reduced levels E-selectin; and(b) the treatment is modified to prevent or treat a hematotoxicity associated with the administration of the cell.

2. A method of predicting or diagnosing hematotoxicity in a subject undergoing treatment with a cell comprising a nucleic acid encoding a CAR that expresses the CAR on its surface, the method comprising:(a) determining that(i) the levels of ANG1, MMP-1 and / or E-selection are reduced in a sample isolated from the subject following administration of the cell;(ii) the levels of thrombomodulin and / or VCAM-1 are increased in a sample isolated from the subject following administration of the cell; and / or(iii) the ratio of ANG2:ANG1 is increased in a sample isolated from the subject following administration of the cell; and(b) classifying the subject as being at risk of or suffering from a hematotoxicity following the determination of (a).

3. A method of treating cancer comprising:(a) administering a therapeutically effective dose of cells comprising a nucleic acid encoding a CAR, wherein the cells express the CAR on their surface;(b) predicting or diagnosing a hematotoxicity in accordance with the method of claim 2; and(c) modifying the treatment to prevent or treat the hematotoxicity associated withthe administration of the cell.

4. The cell for use of claim 1 or the method of claim 3, wherein the treatment is modified by:(i) reducing the dose of cells administered to the subject; and / or(ii) administering a salvage therapy (e.g., a transfusion and / or administering G- CSF).

5. A cell comprising a nucleic acid encoding a CAR that expresses the CAR on its surface for use in a method of treating cancer in a subject, wherein:(i) the VCAM-1 and / or sIL-2R levels of the subject prior to treatment are below a threshold value;(ii) the tumor burden of the subject prior to treatment is below a threshold value; and / or(iii) the percentage of intermediate monocytes in the blood of the subject prior to treatment is below a threshold value; and / or(iv) the ANG1 levels of the subject prior to treatment are above a threshold value.

6. A method for assessing the likelihood that a subject would suffer a severe hematotoxicity if treated with a cell comprising a nucleic acid encoding a CAR that expresses the CAR on its surface, the method comprising:(a) determining that(i) the levels of VCAM-1 and / or sIL-2R in a sample isolated from the subject prior to treatment are below a threshold value;(ii) the tumor burden of the subject prior to treatment is below a threshold value;(iii) the percentage of intermediate monocytes in a blood sample isolated from the subject prior to treatment is below a threshold value; and / or(iv) the levels of ANG1 in a sample isolated from the subject prior to treatment are above a threshold value; and(b) classifying the subject as being at low risk of suffering from a severe hematotoxicity following the determination of (a).

7. A method of treating cancer comprising:(a) determining that the subject is at low risk of suffering from a severe hematotoxicity in accordance with the method of claim 6; and(b) administering a therapeutically effective dose of cells comprising a nucleic acid encoding a CAR, wherein the cells express the CAR on their surface.

8. The cell for use of claim 5 or the method of claim 6 or 7, wherein the VCAM-1 and / or sIL-2R levels are assessed in combination with an EASIX score, mEASIX score or CAR- Hematotox score.

9. The cell for use of claim 5 or 8 or the method of any one of claims 6-8, wherein:(i) the levels of VCAM-1 are below the upper quartile of the patient population suffering from cancer (e.g., suffering the same cancer as the subject);(ii) the levels of sIL-2R are below the upper quartile of the patient population suffering from cancer (e.g., suffering the same cancer as the subject);(iii) the levels of VCAM-1 are below about 2100 ng / ml (e.g., below 2032 ng / ml); and / or(iv) the levels of sIL-2R are below about 5400 pg / ml or below about 5000 pg / ml (e.g., below 4916 pg / ml).

10. The cell for use of any one of claims 1, 4, 5, 8 and 9, or the method of any one of claims 2-4 and 6-9, wherein the sample of the subject that is assessed is a blood or serum sample.

11. The cell for use of any one of claims 1, 4, 5, 8, 9 and 10, or the method of any one of claims 2-4 and 6-10, wherein(i) the CAR targets a cancer antigen (e.g., the CAR is an anti-BCMA or anti-CD19 CAR);(ii) the cell is an immune cell (e.g., a T cell); and / or(ii) the subject suffers from a blood cancer (e.g., lymphoma or multiple myeloma).

12. The cell for use of claim 11 or the method of claim 11, wherein:(i) the cell is idecabtagene vicleucel and / or the subject suffers from multiple myeloma;(ii) the cell is axicabtagene ciloleucel and / or the subject suffers from lymphoma; or(iii) the cell is ciltacabtagene autoleucel and / or the subject suffers from multiplemyeloma.

13. The cell for use of claim 11 or 12, or the method of claim 11 or 12, wherein:(i) the cell is idecabtagene vicleucel and / or the subject suffers from multiple myeloma; or(ii) the cell is axicabtagene ciloleucel and / or the subject suffers from lymphoma.

Images