Method for optimizing t-cell products

Abstract

The invention relates to a method for controlling the quality of T cells during and / or after cultivating said T cells and / or for controlling and optimizing the cultivation of T cells. The method has the steps of: a) isolating at least one nucleic acid molecule of at least one T cell during and / or after the cultivation; b) determining the methylation degree of at least one CpG dinucleotide of the nucleic acid molecule, wherein the CpG dinucleotide is selected from the group consisting of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824, cg13789303, and at least one CpG dinucleotide which is located within the region of 500 nucleotides upstream and / or downstream of each of the aforementioned CpG dinucleotides; and c) comparing the methylation degree determined in step b) with at least one reference value which corresponds to the methylation degree of the CpG dinucleotide of uncultivated primary T cells. The result of the comparison directly indicates whether the cultivated T cells are suitable for therapeutic purposes or if the cultivation conditions should be modified.

Description

[0001] RWTH Aachen 23513WO Method for optimizing T-cell products Background of the invention The invention relates to a method for quality control of T-cell production. In particular, the invention relates to a method for monitoring the quality of T-cells during and / or after culturing these T-cells, and / or for monitoring and optimizing the cultivation of T-cells. State of the art Cellular immunotherapy with autologous, genetically modified T-cells has achieved remarkable initial response rates in various hematological malignancies such as acute lymphocytic leukemia (ALL) or lymphoma (1). After treatment with these chimeric antigen receptor (CAR) T-cells, approximately 30-40% of patients show long-term remission (2). The number of CAR-T cells in a formulated drug is normally between 10 6 and 10 7cells per kilogram of body weight, which requires culture expansion (3). However, little is known about how culture conditions, such as the duration of culture expansion, affect the molecular and functional properties of CAR T cells. A better understanding of the molecular signatures that reflect the therapeutic activity of CAR T cells could help to further optimize the manufacturing process (4). DNA methylation (DNAm) at specific CG dinucleotides (CpG sites) in the genome is one of the most important epigenetic processes. It is mediated by epigenetic writers, such as the de novo methyltransferases DNMT3A and DNMT3B or TET methylcytosine dioxygenases (5). Some fundamental details about the precise regulation of DNAm patterns during development and their functional implications remain to be clarified (6). In any case, DNA methylation varies between cell types,undergoes highly consistent changes during aging and malignant transformation, making them a well-suited biomarker for diagnostic applications (7). In a recent study, the epigenetic landscape of CD19-directed CAR T cells was analyzed prior to infusion. The authors identified DNAm changes caused by transduction of cells with the CAR vector. Their epigenetic signature, termed EPICART, was associated with complete response (CR), event-free survival (EFS), and overall survival (OS) after infusion (8). However, this signature is based on DNAm changes caused by transduction of cells with the CAR vector, so it is largely unclear how the DNAm of CAR T cells changes over time, independent of transduction with the viral CAR vectors.in cell culture. Furthermore, the DNAm changes within the EPICART signature are poorly reproducible. In vitro expansion of stem and progenitor cells generally appears to be associated with a continuous decline in function and eventual cessation of proliferation—a state referred to as replicative senescence. The inventors and others have previously shown that culture expansion of mesenchymal stromal cells (MSCs), fibroblasts, and hematopoietic stem and progenitor cells (HSPCs) is accompanied by highly reproducible DNAm changes, possibly caused by a stochastic process termed "epigenetic drift" (9). Long-term culture-associated DNAm changes could be used to predict time in culture or cumulative population doubling (10). Importantly,that the culture-associated DNAm changes are related to, but not identical to, the so-called "epigenetic clocks" that reflect the aging of the organism (9, 11). Furthermore, there are cell-type-specific differences in DNAm changes during in vitro expansion. Description of the Invention The object of the invention is to provide an improved method for the quality control of cultured T cells that exhibits high specificity and sensitivity, thus enabling a reliable prediction of the therapeutic suitability of the cells as well as a reliable optimization of the cultivation conditions. According to the invention, this object is achieved by a method of the type mentioned above, which comprises the following steps: a) isolating at least one nucleic acid molecule of at least one T cell during and / or after cultivation; b) determining the methylation level of at least one CpG dinucleotide of the nucleic acid molecule,wherein the CpG dinucleotide is selected from the group consisting of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824, cg13789303 and at least one CpG dinucleotide located within a region of 500 nucleotides upstream and / or downstream of each of said CpG dinucleotides; c) comparing the methylation level determined under b) with at least one reference value corresponding to the methylation level of the CpG dinucleotide of uncultured primary T cells, whereby the result of the comparison directly indicates whether the cultured T cells are suitable for therapeutic purposes or whether the cultivation conditions should be changed. The in vitro method according to the invention is based on DNAm changes during the culture expansion of T cells and CAR-T cells. The invention has been able to demonstratethat several CpGs are continuously hypermethylated, particularly within genes related to T cell function, which have previously been linked to T cell exhaustion in vivo. Due to the stochastic pattern of DNAm at neighboring CpGs, the process does not appear to be directly regulated, but rather resembles a stochastic drift. These DNAm changes are indicative not only of the duration of cultivation prior to infusion, but also of long-term survival after transfusion. Thus, not only recombinant modifications of T cells, e.g., through CAR constructs, but also the in vitro production methods can influence the therapeutic efficacy of T cells. The epigenetic signature of the invention, which is based on 14 culture-associated CpGs (Table A),For example, it enables a reliable prediction of cell culture time under different culture conditions. For example, CAR-T cell products with higher DNAm levels at these CpGs are associated with a significantly lower long-term survival rate after transfusion. Surprisingly, it has been found that cell culture expansion of CAR-T cells induces DNA hypermethylation at specific sites in the genome, which apparently reflects epigenetic drift.which is associated with a loss of T cell function and a poorer therapeutic outcome. The method according to the invention for the quality control of cultured T cells exhibits high specificity and sensitivity and enables a reliable prediction of the therapeutic suitability of the T cells as well as a reliable optimization of the cultivation conditions. Table A shows the positions of the 14 culture-associated CpGs according to the invention in the human genome, whereby the sample number "cg..." in the sense of the invention refers to the position of the respective CpG dinucleotide on the "Infinium MethylationEPIC BeadChip" (Illumina, San Diego, California, USA). Table B shows the regions (genome segments) of 500 nucleotides upstream and / or downstream of each of the CpG dinucleotides according to the invention. Table A: Genome Reference Consortium,Human Build 38 Organism: Homo sapiens (human); GRCh38 Sample No. CpG position in the genome* Associated gene cg08364283 chr12: 120248786-120248788 PXN cg03898320 chr2: 207135491-207135493 KLF7 cg20606093 chr14: 68268353-68268355 RAD51B cg21108925 chr2: 28356267-28356269 Not further classified cg07279377 chr2: 201191289-201191291 CASP10 cg14117392 chr5: 139638356-139638358 Not further classified further classified cg04455867 chr2: 181372220-181372222 LOC101927156 cg13298528 chr11: 118893153-118893155 CXCR5 cg06175418 chr5: 134118082-134118084 TCF7 cg04867484 chr4: 148376316-148376318 NR3C2 cg18387515 chr12: 45655453-45655455 Not further classified cg12067423 chr15: 84598401-84598403 Not further classified cg09801824 chr10: 72317171-72317173 Not further classified cg13789303 chr6: 108565477-108565479 FOXO3 (* “chr” = chromosome) Table B: Range + / - 500 nucleotides Genome Reference Consortium,Human Build 38 Organism: Homo sapiens (human); GRCh38 Sample No. CpG position in the genome* cg08364283 chr12: 120248286-120249288 cg03898320 chr2: 207134991-207135993 cg20606093 chr14: 68267853-68268855 cg21108925 chr2: 28355767-28356769 cg07279377 chr2: 201190789-201191791 cg14117392 chr5: 139637856-139638858 cg04455867 chr2: 181371720-181372722 cg13298528 chr11: 118892653-118893655 cg06175418 chr5: 134117582-134118584 cg04867484 chr4: 148375816-148376818 cg18387515 chr12: 45654953-45655955 cg12067423 chr15: 84597901-84598903 cg09801824 chr10: 72316671-72317673 cg13789303 chr6: 108564977-108565979 (* "chr" = chromosome) In an advantageous embodiment of the method according to the invention, the T cells are recombinant T cells that comprise at least one gene encoding a modified antigen receptor. Preferably, the gene is a genethat encodes an adapted T cell receptor (TCR) or a chimeric antigen receptor (CAR). Alternatively, the method according to the invention also encompasses the possibility of isolating T cells that do not contain recombinant genes from tumors, expanding them ex vivo (e.g., in Prodigy), and then reinjecting them ("Adoptive Cell Therapy (ACT)", (56)). Since the specific DNAm changes also occur or could be detected in non-transduced mock cells, the method according to the invention can be applied to all T cells. In a further advantageous embodiment of the method according to the invention, it is provided that at least one CpG dinucleotide is located within a region of 400, 300, 200 or 100 nucleotides upstream and / or downstream of each of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423,cg09801824 and cg13789303. Tables C to F show the corresponding regions (genome segments) of 400, 300, 200, and 100 nucleotides upstream and / or downstream of each of the CpG dinucleotides according to the invention, respectively. Table C: Range + / - 400 nucleotides Genome Reference Consortium,Human Build 38 Organism: Homo sapiens (human); GRCh38 Sample No. CpG position in the genome* cg08364283 chr12: 120248386-120249188 cg03898320 chr2: 207135091-207135893 cg20606093 chr14: 68267953-68268755 cg21108925 chr2: 28355867-28356669 cg07279377 chr2: 201190889-201191691 cg14117392 chr5: 139637956-139638758 cg04455867 chr2: 181371820-181372622 cg13298528 chr11: 118892753-118893555 cg06175418 chr5: 134117682-134118484 cg04867484 chr4: 148375916-148376718 cg18387515 chr12: 45655053-45655855 cg12067423 chr15: 84598001-84598803 cg09801824 chr10: 72316771-72317573 cg13789303 chr6: 108565077-108565879 (* “chr” = chromosome) Table D: Range + / - 300 nucleotides Genome Reference Consortium,Human Build 38 Organism: Homo sapiens (human); GRCh38 Sample No. CpG position in the genome* cg08364283 chr12: 120248486-120249088 cg03898320 chr2: 207135191-207135793 cg20606093 chr14: 68268053-68268655 cg21108925 chr2: 28355967-28356569 cg07279377 chr2: 201190989-201191591 cg14117392 chr5: 139638056-139638658 cg04455867 chr2: 181371920-181372522 cg13298528 chr11: 118892853-118893455 cg06175418 chr5: 134117782-134118384 cg04867484 chr4: 148376016-148376618 cg18387515 chr12: 45655153-45655755 cg12067423 chr15: 84598101-84598703 cg09801824 chr10: 72316871-72317473 cg13789303 chr6: 108565177-108565779 (* “chr” = chromosome) Table E: Range + / - 200 nucleotides Genome Reference Consortium,Human Build 38 Organism: Homo sapiens (human); GRCh38 Sample No. CpG position in the genome* cg08364283 chr12: 120248586-120248988 cg03898320 chr2: 207135291-207135693 cg20606093 chr14: 68268153-68268555 cg21108925 chr2: 28356067-28356469 cg07279377 chr2: 201191089-201191491 cg14117392 chr5: 139638156-139638558 cg04455867 chr2: 181372020-181372422 cg13298528 chr11: 118892953-118893355 cg06175418 chr5: 134117882-134118284 cg04867484 chr4: 148376116-148376518 cg18387515 chr12: 45655253-45655655 cg12067423 chr15: 84598201-84598603 cg09801824 chr10: 72316971-72317373 cg13789303 chr6: 108565277-108565679 (* "chr" = chromosome) Table F: Range + / - 100 nucleotides Genome Reference Consortium, Human Build 38 Organism: Homo sapiens (human); GRCh38 Sample No. Position in the genome* cg08364283, 120248686-120248888 cg03898320 chr2: 207135391-207135593 cg20606093 chr14: 68268253-68268455 cg21108925 chr2: 28356167-28356369 cg07279377 chr2: 201191189-201191391 cg14117392 chr5: 139638256-139638458 cg04455867 chr2: 181372120-181372322 cg13298528 chr11: 118893053-118893255 cg06175418 chr5: 134117982-134118184 cg04867484 chr4: 148376216-148376418 cg18387515 chr12: 45655353-45655555 cg12067423 chr15: 84598301-84598503 cg09801824 chr10: 72317071-72317273 cg13789303 chr6: 108565377-108565579 (* “chr” = chromosome) In a further advantageous embodiment of the invention The procedure provides that suitability for therapeutic purposes is assumed if the methylation level determined under b) corresponds essentially to the reference value, or that suitability for therapeutic purposes is not assumed if the methylation level determined under b) is increased compared to the reference value.In a further advantageous embodiment of the method according to the invention, it is provided that the cultivation conditions can remain unchanged if the degree of methylation determined under b) essentially corresponds to the reference value, or the cultivation conditions should be changed if the degree of methylation determined under b) is increased compared to the reference value. The cultivation conditions can advantageously be changed such that the changed conditions include a change in the composition of the culture medium. Alternatively or additionally, the cultivation conditions can advantageously be changed such that the changed conditions include a shortening of the cultivation time.Alternatively or additionally, the cultivation conditions can advantageously be further modified such that the methylation of the CpG dinucleotide is inhibited, for example by means of inhibitors of DNA methyltransferases, or genetic or epigenetic inactivations of DNA methyltransferases (in particular DNMT3A). The object is furthermore also achieved by the use of at least one nucleic acid molecule which comprises at least one of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824 and cg13789303 and / or at least one CpG dinucleotide which is located within a region of 500 nucleotides upstream and / or downstream of each of the said CpG dinucleotides, for controlling the Quality of T cells during and / or after cultivation of these T cells.In an advantageous embodiment of this use according to the invention, the T cells are recombinant T cells comprising at least one gene encoding a modified antigen receptor. Preferably, the gene is a gene encoding an adapted T cell receptor (TCR) or a chimeric antigen receptor (CAR). The object is further achieved by the use of at least one nucleic acid molecule which comprises at least one of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824 and cg13789303 and / or at least one CpG dinucleotide which is located within a region of 500 nucleotides upstream and / or downstream of one of the said CpG dinucleotides, for the control and optimization a method for culturing T cells.In an advantageous embodiment of this use according to the invention, the T cells are recombinant T cells that comprise at least one gene encoding a modified antigen receptor. Preferably, the gene is a gene encoding an adapted T cell receptor (TCR) or a chimeric antigen receptor (CAR). The degree of methylation or measurement of DNA methylation (DNAm) can be carried out, for example, using pyrosequencing or MassArray, cost-effectively and in a high-throughput process for the specific regions (e.g., DNA sections). Alternatively, more advanced examination techniques such as sequencing and microarray-based methods are also possible, for example, digital droplet PCR, Epityper, amplicon sequencing, nanopore sequencing, SMART sequencing, and the like.The invention further comprises, for example, a kit with a compilation of suitable reference DNAs, primer sets, as well as optional instructions and evaluation software. The object is thus also achieved by a kit, in particular a kit for monitoring the quality of T cells during and / or after culturing these T cells, and / or for monitoring and optimizing the cultivation of T cells, preferably for carrying out the method according to the invention, which comprises at least one oligonucleotide primer (for example, oligonucleotide-primer pair, e.g.a “primer set”, see Table S3 below) for amplifying and / or sequencing at least one CpG dinucleotide of at least one nucleotide sequence comprising at least one CpG dinucleotide from the group consisting of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824, cg13789303 and at least one CpG dinucleotide located within a region of 500 nucleotides upstream and / or downstream of each of said CpG dinucleotides, and comprises at least one reference nucleic acid.The kit according to the invention can, for example, comprise at least one artificial reference nucleic acid molecule according to the invention and optionally at least one buffer solution and / or at least one reagent for carrying out at least one method selected from the group consisting of DNA amplification, bisulfite treatment of DNA, DNA sequencing, preferably pyrosequencing of DNA, MassArray analysis, deep sequencing of bisulfite-converted DNA, flow cytometry-based bead assays, SNP genotyping, digital droplet PCR, Epityper, amplicon sequencing, nanopore sequencing and SMART sequencing. The invention is explained in more detail below with reference to the figures, tables and examples. Brief description of the figures Figure 1: CAR-T cells accumulate DNA methylation changes during culture expansion. a.Different culture conditions for the expansion of non-transduced T cells (UT) and CAR T cells for up to 22 days. b. Growth curves of UT and CAR T cells during small-scale in vitro expansion. The arrows indicate the (re)stimulation of the T cells with particles. c. Principal component analysis (PCA) of the DNA methylation profiles. The arrows indicate trajectories with culture time (blue shades) for the different culture conditions. d. Volcano plot showing differentially methylated CpG sites (hypermethylated in red, hypomethylated in blue) in d0 (UT) versus d22 samples (UT and CAR). e. Bubble plots of the top 10 Gene Ontology (GO) terms for significantly hyper- and hypomethylated CpGs after 22 days of culture expansion (gene ratio = differentially methylated / total number of genes in this pathway; bubble size = absolute number of genes in the set; color code indicates significance level). f.The scattergrams compare the mean methylation changes (mean Δβ value) during in vitro expansion in T cells (UT and CAR; d0 versus d22) compared to culture of hematopoietic stem and progenitor cells (HSPCs), human umbilical vein endothelial cells (HUVECs), and mesenchymal stromal cells (MSCs), respectively. The percentages of CpG sites with more than 10% mean DNAm change are highlighted (red = hypermethylated in both cell types; blue = hypomethylated in both cell types; black = divergently methylated; r = Pearson correlation coefficient). All CpGs shared by the 450K and EPIC platforms are shown here. Figure 2: Culture expansion of T cells leads to downregulation of genes relevant for T cell function. a. Mean plot (MA) of RNA-seq data illustrating differentially expressed genes in T cells after 22 days of culture expansion (adjusted p < 0.05, n = 4). b.Correlation of gene expression changes with DNA methylation changes during culture expansion (day 0 versus day 22) for CpGs located in promoter regions (left) or in the gene body (right) of the corresponding genes. If multiple CpGs occurred in these regions, the mean of the DNAm was calculated to obtain a single point per gene. c. Gene Ontology classification of differential gene expression during T cell expansion, presented as bubble plots of the top 30 GO terms, ordered by the gene ratio for significantly up- and downregulated genes in T cells after 22 days of culture expansion. Among the downregulated genes, four categories are highlighted in red, which also show the highest enrichment for hypermethylated CpGs (see Fig. 1e). d.Heatmap of gene expression levels (expressed as log counts per million; log CPM) of genes that are significantly downregulated during long-term culture and belong to these four Gene Ontology categories. Figure 3: Neighboring CpGs are stochastically methylated during culture expansion. a. Culture-associated DNA methylation (DNAm) patterns were analyzed in genomic regions of three genes (TOX, SMAD3, and GRAP2) using bisulfite amplicon sequencing (BS-Seq). The heatmaps represent the sequence of methylated and unmethylated sites in individual reads of the three amplicons. The results are shown for a sample on day 0 and day 22. b. Pearson correlation of DNAm at adjacent CpG sites within the amplicons of TOX, SMAD3, and GRAP2 for T cells expanded under two culture conditions (orange: IL2 / particle, n = 8; purple: IL-7 / IL15 / nanomatrix, n = 5).The position of the CpG sites within the amplicons is shown (asterisks mark the CpG with the highest Pearson correlation). c. The linear correlation of DNAm with culture time is shown as an example for the CpGs with the highest correlation. Figure 4: Epigenetic predictions for the time of expansion of CAR T cells in culture. a. Heatmap showing the DNA methylation values ​​(DNAm) of the 339 CpGs with almost linear increase (Pearson correlation r > 0.9; 336 CpGs) or loss of methylation (r < -0.9; 3 CpGs) with time in culture for various T cell preparations in the training set and an independent validation set. The DNAm values ​​(β values) are indicated by the color code. b. Epigenetic predictions based on 35 CpGs (Elastic-Net) correlate with time in culture for the training dataset (r2 = 0.99; n = 15) and the validation dataset (r2 = 0.90; n = 13). c.The epigenetic 35-CpG predictor of time in culture was subsequently applied to DNAm profiles from three clinical trials: NCT02772198 (n = 43), NCT03144583 (n = 45), and NCT03373071 (n = 26). The interleukins delivered during expansion and the actual time of culture (shaded areas) are indicated. d. The DNAm profiles from these three studies were subsequently stratified into those with or without subsequent onset of cytokine release syndrome (CRS) or immune effector cell-associated neurotoxicity syndrome (ICANS). The boxplot indicates the predicted time in culture (interquartile range (IQR) and whiskers denote the 1.5 × IQR; p calculated using Student's t-test). e. The association of DNA methylation levels in individual culture-associated CpGs of CAR T cells with overall survival after transplantation was tested.The scatterplot shows the hazard ratios and p-values ​​for the 339 culture-associated CpGs (adjusted for disease and clinical trial; not adjusted for multiple testing of 339 CpGs) in the test dataset from three clinical trials (GSE179414; n = 114). Yellow dots mark the 35 CpGs of the elastic net predictor for time in culture. Figure 5: Culture-associated DNA methylation changes in CAR T cells are an indicator of treatment success. a. The association of culture time-associated DNA methylation changes (DNAm) with overall survival was tested in a target identification cohort comprising three clinical trials (subset of GSE179414; n = 82). Multivariate Cox regression analysis revealed that of the 339 CpGs with the highest correlation with culture time, 14 CpG sites were associated with overall survival (adjusted for disease and clinical trial; not adjusted for multiple testing of 339 CpGs). b.For these 14 CpGs, the association with overall survival was further tested for the target identification cohort (orange; n = 82) and the target validation cohort (turquoise; n = 32). Hazard ratios (adjusted for disease and clinical trial) and confidence intervals (CI) are shown (*p < 0.01). c. Subsequently, a multivariable epigenetic predictor for culture time based on the 14 CpGs was trained (r2 = 0.94; n = 15) and validated (r2 = 0.70; n = 13) in independent datasets (as shown in Fig. 4a). d. Cox proportional hazards model (adjusted for disease and clinical trial) for the epigenetic prediction of time in culture based on the 14 CpGs. A longer predicted time in culture of CAR T cells was associated with a higher risk of death (*p-value < 0.01; ***p < 0.0001). e. Kaplan-Meier estimates of overall survival for the target identification (n = 82) and validation cohorts (n = 32).Subgroups were divided into low and high estimates for culture-associated DNAm based on the 14-CpG predictor. f. Boxplot showing the comparison of predicted time in culture between patient cohorts without / with cytokine release syndrome (CRS) or immune effector cell-associated neurotoxicity syndrome (ICANS). The boxes indicate the interquartile range (IQR) and the whiskers the 1.5 × IQR (*p < 0.01). Figure S1: Immunophenotypic analysis during small-scale expansion for up to 22 days. a. Representative gating strategy for T cells (CD3+); CD4+ and CD8+ subpopulations; Naive: CD62L+ , CD95- and CD45RO- ; Stem cell memory (TSCM): CD62L+, CD95+, and CD45RO-; central memory (TCM): CD62L+, CD95+, and CD45RO+; effector memory (TEF): CD62L-, CD95+, and CD45RO+; effector (TEFF): CD62L-, CD95+, and CD45RO-. b.Changes in the proportion of different T cell subsets during culture expansion (with IL-7 / IL-15 / nanomatrix; blue = untransduced (UT) T cells; purple = CAR T cells; n = 3). c. Example changes in the ratio between CD4+ and CD8+ T cells expanded under the same culture conditions with IL-7 / IL-15-saturated medium (n = 3). A similar shift in favor of CD8+ T cells was observed with IL-2-saturated culture medium (data not shown). d. Representative gating strategy for the analysis of CAR surface expression in transduced T cells. e. Summary of transduced T cells expressing CAR on their surface at different time points during culture expansion (means ± SEM). Figure S2: Filtering strategy and analysis of DNA methylation changes a.Predicted percentage of blood cell types estimated by the Houseman algorithm1 in EPIC BeadChip datasets of T cells from healthy donors; with or without CAR transduction, cultured with two different culture regimes and for different times (up to 22 days). The DNA methylation (DNAm) profiles were correctly assigned as T cells with decreasing CD4 / CD8++ ratios by analogy with immunophenotypic analysis. b. To reduce the effects of cellular composition, we exclude CpGs with significant DNAm differences between CD4+ and CD8+ T cells. To identify these CpGs, we used a public dataset (GSE110554)2 to select CpGs with a mean DNAm difference >10% or an adjusted limma p <0.05; 46856 CpGs; purple dots). c. CpG filtering strategy for further analysis of culture-associated DNAm alterations.Of the CpGs represented by the EPIC BeadChip, the inventors identified CpGs with SNPs, cross-reactive probes3,4, autosomes, and the 46,856 CpGs with different DNAm between CD4+ and CD8+ T cells. d. Principal component 2 in the PCA of the DNAm profiles in Fig. 1c correlates with time in culture. e. Pairwise comparison of mean DNAm differences between untransduced (UT) and CAR T cells on day 22 (n = 3; all expanded with IL-7 / IL-15 / nanomatrix). There were no significant differences (limma-adjusted p < 0.05). f. Scatterplot of DNAm differences between untransduced T cells cultured with either IL-2 / particles or IL-7 / IL-15 / nanomatrix for 22 days. The number of significant CpGs is indicated in purple and yellow, respectively (mean difference in DNAm level > 20% and limma-adjusted p < 0.05). Figure S3: Comparison of age- and culture-related DNAm changes in T cells.To further investigate whether culture-associated DNAm alterations are related to age-associated DNAm alterations in T cells, we used DNAm profiles of flow-sorted CD4 cells+ from healthy controls (GSE67705; n = 24; age range: 26–66 years). This dataset was analyzed with the 450K BeadChip, and therefore we focused on the 386,700 CpGs shared with the EPIC platform. Of 339 culture-associated CpGs, a total of 138 CpGs were also represented by the 450K BeadChip and are highlighted in red. Figure S4: Relationship between the 35CpG predictor and clinical data. a. Scatterplot showing that epigenetic predictions of time in culture using the 35 CpG Elastic Net predictor have little correlation with chronological age in the validation dataset (from Fig. 4a; n = 12) and the additional test dataset (GSE179414 from Fig. 4c; n = 114).Blue dots = CAR T cells for patients with acute lymphoblastic leukemia (ALL); purple dots = CAR T cells for patients with non-Hodgkin lymphoma (NHL). b. Epigenetic predictions of time in culture based on the 35-CpG elastic network model were stratified by disease (ALL versus NHL). Figure S5: Kaplan-Meier estimates of overall survival for DNAm at five culture-associated CpG sites. Multivariate Cox regression analysis of the target identification (a; n = 82) and validation cohorts (b, n = 32) showed that five CpGs associated with culture time were associated with a higher death rate in the independent target validation cohort. Kaplan-Meier estimates for these subcohorts, where patients were divided into quartiles based on DNAm values, are presented here. Wald test p-values ​​are shown.Figure S6: Association between epigenetic predictions and overall survival in individual clinical trials. a. Stacked bar chart of patient distribution per clinical trial according to predicted time in culture, categorized into quartiles such as short, medium-short, medium-long, and long predicted time in culture. The longer time in culture of NCT03373071 (14 days) is also reflected in higher estimates in the upper quartile (for NCT02772198 and NCT03144583, no precise information on the exact culture time for each sample is available). b. Kaplan-Meier estimates of overall survival comparing the three clinical trials. There were no significant differences in outcome between these trials (shown here as an example compared to NCT02772198). c. Kaplan-Meier estimates of overall survival for short and long predicted time in culture for each individual clinical trial (log-rank p-values ​​are given).Description of exemplary and advantageous embodiments of the invention Chimeric antigen receptor T cells (CAR cells) offer new perspectives for the treatment of hematological malignancies. The production of these cellular products involves culture expansion processes that can compromise cellular integrity and therapeutic outcome. In this study, the inventors investigated culture-associated epigenetic changes in CAR T cells and observed a continuous increase in DNAm, particularly in genes relevant to T cell function. Hypermethylation in many genes, such as TCF7, RUNX1, and TOX, was reflected in transcriptional downregulation. 339 CG dinucleotides (CpGs) showed an almost linear increase in methylation over time in cell culture, although neighboring CpGs were not coherently regulated on the same DNA strands.An epigenetic signature based on 14 of these culture-associated CpGs reliably predicted cell culture time under various culture conditions. Notably, CAR T cell products with higher DNAm levels at these CpGs were associated with significantly lower long-term post-transfusion survival. The data demonstrate that cell culture expansion of CAR T cells induces DNA hypermethylation at specific sites in the genome, apparently reflecting epigenetic drift associated with a loss of T cell function. Therefore, shorter culture times are advantageous to avoid dysfunctional methylation programs, which appear to be associated with poorer therapeutic outcomes.Materials and Methods Ethical Approval and Consent to Participation All blood donors and patients provided written informed consent in accordance with the Declaration of Helsinki and the International Conference on Harmonization guidelines for good clinical practice. The isolation of T cells from the peripheral blood of healthy donors was performed according to the guidelines approved by the local ethics committees of RWTH Aachen University (EK 041 / 15). The clinical studies NCT03853616 and NCT03870945 were approved by the Paul Ehrlich Institute (Germany). All blood samples were handled in compliance with the required ethical and safety procedures. Blood Samples T cells were isolated from the peripheral blood of healthy donors. For the generation of CAR T cells, either buffy coats or leukapheresis products from healthy donors were used, which were provided by the German Red Cross Dortmund and the German Red Cross Dortmund, respectively.from the German Red Cross (DRK) Ulm and the Hannover Medical School (MHH). Furthermore, the inventors analyzed CAR-T cells obtained from fresh, non-mobilized leukapheresis products from patients with acute lymphocytic leukemia (ALL) or diffuse large B-cell lymphoma (DLBCL) participating in the MB-CART19.1 or MB-CART2019.1 clinical trials (NCT03853616 and NCT03870945). Cell culture: In this study, the inventors compared different cell culture methods: For the expansion of T cells with CD2 / CD3 / CD28-loaded particles, the inventors isolated peripheral blood mononuclear cells (PBMCs), selected CD3+ cells using the MicroBead Kit (Miltenyi Biotec), and confirmed successful enrichment by flow cytometry.T cells (106 cells / ml) were resuspended in RPMI (Gibco Thermo Fisher), 10% FCS (Gibco Thermo Fisher), and 20 U of IL-2 per ml (Miltenyi Biotec) using the T Cell Activation / Expansion Kit, human (with CD2 / CD3 / CD28-loaded MACSiBead particles; Miltenyi Biotec) according to the manufacturer's instructions. Restimulation was performed on day 7 and day 17. At each passage, cells were counted using a Neubauer chamber to calculate cumulative population doublings. For T cell / CAR-T cell expansion with CD3 / CD28-conjugated polymeric nanomatrix, we selected T cells using the human Pan T Cell Isolation Kit (Miltenyi Biotec) and validated the enrichment by flow cytometry.Cells were resuspended in TexMACS™ medium (Miltenyi Biotec) supplemented with recombinant human IL-7 (12.5 ng / mL) and IL-15 (12.5 ng / mL, both Miltenyi Biotec) and activated with a polymeric nanomatrix conjugated to humanized CD3 and CD28 (T Cell TransAct™; Miltenyi Biotec). For automated T cell expansion, all steps of cell enrichment, lentiviral CAR transduction of T cells, as well as cultivation and media exchange, were performed using the CliniMACS Prodigy® (Miltenyi Biotec). Specifically, CD4+ and CD8+ T cells were enriched from leukapheresis products of healthy donors and patients by magnetic positive selection, directly followed by T cell activation with T Cell TransAct™ (Miltenyi Biotec). After 24 hours, T cells were transduced with a lentiviral vector expressing either a CAR against human CD19 or a tandem CAR against human CD20 and CD19 (lentigen).The T cells were expanded in TexMACS GMP medium supplemented with IL-7, IL-15, and 3% human AB serum at 5% CO2 and 37°C. After 5 days, the medium was replaced with TexMACS GMP medium supplemented with IL-7 and IL-15. Cells were harvested and formulated after 12 days using Composol or CliniMACS formulation solution supplemented with 2.5% human AB serum (HSA). Preparation of CAR constructs and lentiviral vectors: For small-scale manual culture experiments, the inventors used the anti-LLE biotin adapter CAR (AdCAR). Briefly, an anti-LLE scFv was cloned onto a third-generation backbone. The scFv was preceded by an IL-22 leader sequence, followed by a CD8α hinge and a transmembrane domain. The intracellular domains consisted of CD28, followed by 4-1BB as costimulatory domains and a CD3ζ main stimulatory domain.The lentiviral particles were resuspended in TexMACS and stored at -70°C until further use. T cells were transduced 24 hours after activation, and culture was continued according to the recommended T Cell TransAct™ protocol. Flow cytometry: For immunophenotypic analysis of the initial cell preparations and during culture expansion, cells were stained against CD3 (clone REA613), CD4 (clone REA623), CD8 (REA734), CD45RO (clone REA611), CD62L (clone REA615), CD95 (clone REA738), and CAR (anti-biotin-PE clone REA746; all Miltenyi Biotec) (concentration 1:50; 10 min, at 4°C). After incubation, the suspension was washed with AutoMACS Running Buffer and fluorescence was measured using a MACSQuant Analyzer 10 and MACSQuantify™ Software v2.13.0.DNA methylation profiling. Genomic DNA (gDNA) was isolated using the NucleoSpin Tissue Kit (Macherey-Nagel) or the DNeasy Blood & Tissue Kit (Qiagen), converted with bisulfite, and hybridized to the Infinium MethylationEPIC BeadChips (Illumina, San Diego, California, USA) at Life and Brain GmbH (Bonn, Germany). Raw data were preprocessed using the Bioconductor Illumina Minfi package for R and normalized with ssNoob (12). For further analysis, we excluded CpGs with common single nuclear polymorphisms (SNPs; MAF > 1%), CpGs of X and Y chromosomes, and probes identified as cross-reactive. Since the CD4 / CD8++ ratio changes during culture expansion, the inventors also filtered for CpGs that were significantly differently methylated between purified CD4+ and CD8+ cells (GSE110554): 46.856 CpGs with a mean DNAm difference of >10% or an adjusted p < 0.05 (Limma t-test) between CD4+ and CD8+ cells were excluded. Principal component analysis was performed in R and plotted using the ggplot2 package version 3.3.6. Volcano plots were generated using the EnhancedVolcano R package version 1.12.0. Gene ontology analysis was performed for hypo- and hypermethylated CpG sites (DNAm change in T cells on day 0 versus day 22 in culture of more than 20% and adjusted p < 0.05) in R using the missMethyl package. Heatmaps were generated using the ComplexHeatmap R package. Information from the Illumina BeadChip annotation file was used to define promoter categories (TSS1500, TSS200, and 5'UTR) and gene body regions (body, 1st exon, and 3'UTR).To assess whether culture-associated DNAm alterations in T cells are similar in other cell types, we used available datasets from CD34+ cells from cord blood (13), HUVECs (14), and mesenchymal stem cells (15), although these cells were cultured for different lengths of time and under different culture conditions. Since all these datasets were generated on 450K Illumina BeadChips, we focused on CpGs common to the 450K and EPIC platforms, although batch variations between studies and platforms cannot be excluded. Epigenetic predictors of time in culture and clinical outcome Culture-associated CpGs were initially selected based on the Pearson correlation of DNAm values ​​(β values) and days in culture at a threshold of r > 0.9 and r < -0.9 (336 hyper- and 3 hypomethylated CpGs, respectively).To exclude age-associated CpGs, which might be influenced by donor age rather than culture expansion, we used a dataset of peripheral white blood cells (GSE115278; n = 474; mean age: 47.2 ± 14.1 years). Here, 7 of 339 CpGs correlated with chronological age (r > 0.3 or r < -0.3) and were removed. Time-in-culture predictors were then designed as linear regression models fitted to either the elastic net regularization path or a rigid path in the training dataset using the R package glmnet (16). For the elastic net regression model, the mixture parameter was set to α = 0.5. Best-fitting candidate CpGs and their coefficients were selected by 10-fold cross-validation (Supplementary Table S1).For the rigid regression model, candidate CpGs with an association with overall survival were selected from a test dataset for target identification (subset of GSE179414) (8). For this purpose, the Cox proportional hazards regression model was applied using the R package survival version 3.4-0. CpG sites with a p < 0.01 were selected as candidates for the rigid regression model. The coefficients were selected by 10-fold cross-validation on the predictor training dataset (Supplementary Table S2). Survival curves were plotted using the R package survminer version 0.4.9. The significance level is indicated by p-values ​​calculated using a Wald test based on the Cox model. Analysis of gene expression Total RNA was isolated from freshly isolated T cells (day 0) and after three weeks of culture (day 22) in IL-2-supplemented medium using the NucleoSpin RNA Plus Kit (Macherey-Nagel).Library preparation using the QuantSeq 3' mRNA-Seq Library Prep Kit FWD (Lexogen, Vienna, Austria) and sequencing on the HiSeq 2500v4 platform (Illumina; 10 M 50 bp single-end reads) was performed at Life and Brain GmbH. RNA sequencing (RNA-seq) was aligned to the human genome (hg19) using the R package QuasR and Rhisat2 as an aligner with default parameters (17, 18). Reads that aligned with fewer than 10 reads per gene were excluded from further analysis. The number of reads per million (CPM) was calculated and TMMwsp normalized with default parameters using the R package edgeR (19). Count data were voom-transformed using the R package limma (20). For differential expression analysis, the limma model was fitted, and t-statistics were calculated using the empirical Bayesian method with an adjusted p-value of 0.05. Mean-difference plots were generated using limma in R.Gene ontology analysis was performed using the BH method within the R package clusterProfiler. Bisulfite amplicon sequencing: To further investigate DNAm alterations in genomic regions with prominent culture-associated CpGs, we used barcoded bisulfite amplicon sequencing (BA-seq) (11, 14). gDNA was isolated from 2x106 T cells using the DNeasy Blood & Tissue Kit (Qiagen), quantified using a Nanodrop 2000 spectrophotometer (Thermo Scientific), and a total of 200–500 ng of genomic DNA was used for bisulfite conversion using the EZ DNA Methylation Kit (Zymo Research). The target sequences in the genomic regions of TOX, SMAD3, and GRAP2 were amplified using the PyroMark PCR kit (Qiagen) with primers containing a handle sequence (Supplementary Table S3).The PCR products were combined in an equimolar composition, followed by a purification step using the Select-a-Size DNA Clean & Concentrator Kit (Zymo Research) and diluted with 15% PhiX spike-in control to obtain a 20 picomolar DNA library, which was then sequenced on a MiSeq system using Miseq reagents (Illumina) with 250 bp paired-end sequences. The Bismark tool was used to map bisulfite-converted sequence reads and determine DNAm levels for each CpG (21). On average, 1,000 reads per genomic region could be aligned in one sample. Genomic coordinates for hg19, including reference genes and CpG sites, were obtained from UCSC and plotted using the R package Gviz. Results: T cell cultivation is associated with discrete DNA hypermethylation.To gain insight into the epigenetic changes during CAR T cell expansion independent of the specific manufacturing conditions, the inventors investigated different cell culture regimes for T cell expansion with or without transduction with the CAR vector for up to 22 days (Fig. 1a). Expansion with IL-2 and repeated stimulation with CD2 / CD3 / CD28-loaded particles enabled a higher population doubling but was also associated with increased cell death, especially after the second restimulation, which is consistent with previous results (22). In contrast, stimulation of cells with IL-7, IL-15, and a polymeric nanomatrix conjugated with CD3 / CD28 enabled nearly linear expansion rates with a doubling time of approximately 3.3 days (Fig. 1b).Flow cytometric analysis revealed an immunophenotypic shift toward a central memory phenotype (TCM; CD62L+ CD95+ CD45RO+ ), particularly between day 8 and day 15, and an increasing stem cell memory phenotype (TSCM; CD62L+ CD95+ CD45RO- ), independent of CAR transduction (Fig. S1a,b). Notably, the ratio of CD4+ to CD8+ T cells decreased (Fig. S1c), while the proportion of cells with CAR surface expression remained relatively stable during culture expansion of CAR T cells (Fig. S1d,e). Overall, these data illustrate substantial immunophenotypic changes during culture expansion. Subsequently, we analyzed the DNAm profiles of non-transduced T cells (UT) and CAR T cells using Illumina EPIC Bead Chips.Epigenetic deconvolution of leukocyte subsets was consistent with flow cytometric measurements, showing a decrease in the ratio of CD4+ to CD8+ T cells during culture (Fig. S2a). For further analysis, we therefore excluded CpGs that were significantly differentially methylated between CD4+ and CD8+ T cells to avoid cell type bias (adjusted p < 0.05 or DNAm difference > 10%; Fig. S2b-c). Unsupervised principal component analysis (PCA) revealed continuous changes in the DNAm pattern during culture (Fig. 1c). The second component (PC2) even showed a linear correlation with the cumulative number of population doublings (cPD; Fig. S2d). A pairwise comparison of UT and CAR T cells revealed no significant DNAm changes attributable to CAR transduction (adjusted p < 0.05; Fig. S2e).However, significant epigenetic differences were observed between the different culture conditions: 1121 CpGs were hypermethylated in T cells expanded with IL-2, and 90 CpGs were hypomethylated compared to IL-7 / IL-15 (mean DNAm difference > 20% and adjusted p < 0.05; Fig. S2f). When comparing DNAm changes between preparations from day 0 (d0) and day 22 (d22), the differences were even more pronounced: 3128 hypermethylated and 411 hypomethylated CpG sites (Fig. 1d). These CpGs were associated with 1514 and 215 different genes, respectively. Analysis of gene set enrichment showed that culture time-associated hypermethylation was highly enriched in functional categories of leukocyte and T cell activation, whereas hypomethylation was enriched in categories for T cell and leukocyte differentiation (Fig. 1e).To understand whether the epigenetic changes during T cell expansion are cell type-specific, the inventors compared them with culture-associated DNAm from hematopoietic stem and progenitor cells (HSPCs) (13), human umbilical vein endothelial cells (HUVECs) (14), or mesenchymal stem cells (MSCs) (15). Despite the different culture times, culture conditions, and analysis platforms, there was a correlation of culture-associated DNAm changes in HSPCs and CAR-T cells (Pearson correlation coefficient r = 0.3), whereas this was hardly observed in the other, less related cell types (Fig. 1f). Overall, culture expansion of T and CAR-T cells is associated with continuous and pronounced DNAm changes, particularly in genes relevant to T cell function.T cell expansion impairs T cell regulatory networks. The inventors then investigated whether culture-associated DNAm is also reflected in differential gene expression. RNA sequencing data from d0 versus d22 revealed 3154 up- and 2862 downregulated genes in T cells expanded with IL-2 (Fig. 2a; adjusted p < 0.05). Overall, the inventors could not detect a clear association between gene expression and DNAm alterations for CpGs located either in the promoter or in the gene body (Fig. 2b), supporting the notion that an increase in DNAm does not necessarily lead to a downregulation of gene expression (23). Nevertheless, the Gene Ontology (GO) analysis revealed a significant enrichment of downregulated genes in the categories of T cell activation, lymphocyte differentiation, T cell differentiation, and alpha-beta T cell activation (Fig. 2c).Thus, the differentially expressed genes were enriched in very similar functional categories as observed for DNAm alterations. We further investigated the relevant genes in these overlapping pathways (Fig. 2d). Downregulated genes included TCF1 and LEF1, two important regulators of stem cell and memory formation (24). Likewise, the two chemokine receptors IL6R and IL7R, which are involved in the stimulation of naive T progenitor cells (25), were downregulated. Furthermore, they are crucial mediators of anti-tumor activity as well as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), the most frequently observed immune-mediated adverse events following CAR-T cell administration (26).DNA hypermethylation and downregulation of gene expression were also observed for TOX, T-BET, and EOMES, key transcription factors associated with T cell exhaustion in vivo (27). Culture-induced epigenetic drift in differentially methylated regions. To gain better insight into DNAm changes at culture-time-associated regions, we performed targeted bisulfite amplicon sequencing (BS-Seq) for TOX, SMAD3, and GRAP2. If the DNAm changes were mediated by a targeted regulatory protein complex with an epigenetic writer, this would likely also coherently alter neighboring CpGs. In contrast, we observed that the sequence of methylated or unmethylated CpGs within individual reads of the BS-Seq data appeared to be random (Fig. 3a).The stochastic modification therefore suggests that the culture-associated DNAm in CAR-T cells is more likely due to epigenetic drift (9, 11). On the other hand, the average DNAm values ​​in different preparations of T cells and CAR-T cells showed a similar increase at neighboring CpGs (Fig. 3b). Indeed, the topmost CpGs showed an almost linear increase with time in culture, and this was slightly accelerated in T cells with IL-2 culture conditions (Fig. 3c), which corresponds to the higher cumulative population doublings, as stated above. Epigenetic predictor of time in culture of CAR-T cells. Since several CpGs showed almost linear DNAm changes with culture duration, the inventors subsequently created a robust epigenetic signature that could be aligned with culture expansion.For this purpose, the inventors compiled a training dataset consisting of untransduced T cells and CAR T cells with different culture media and manual or automated expansion. The inventors then filtered for CpGs with a very high linear correlation between DNAm and time in culture (44 hybridizations in total; Fig. 4a). This analysis revealed 336 hypermethylated and only 3 hypomethylated CpGs (Pearson correlation r > 0.9 and r < -0.9, respectively). It has previously been shown that culture-associated CpGs can overlap with age-associated CpGs (28, 29). To create an epigenetic predictor of time in culture of T cells, the inventors removed 7 CpGs that showed a correlation with chronological age in blood samples (GSE115278, r > 0.3 or r < -0.3).This overlap was relatively small, and the inventors therefore further tested culture-associated DNAm changes using another dataset of CD4+ T cells of different donor ages (30). This confirmed the assumption that culture-associated DNAm changes in T cells are rather independent of donor age (Fig. S3). Subsequently, the inventors calculated a multivariable linear elastic net regression model with 10-fold cross-validation and derived a predictor for time in culture based on 35 hypermethylated CpGs (Supplementary Table S1). This predictor showed a high correlation with the days in culture of T cells and CAR T cells in the training dataset (r2 = 0.99) and in the independent validation set of clinical CAR T cells (r2 = 0.90; Fig. 4b).Furthermore, the predictions were overall consistent with culture times from publicly available datasets of CAR T-cell products expanded for 9–10 days with IL-2 in NCT02772198 (n = 43), 7–10 days with IL-7 / IL-15 in NCT03144583 (n = 45), or 14 days with IL-7 / IL-15 in NCT03373071 (n = 26; GSE179414; Fig. 4c) (8). These predictions were barely affected by chronological age or disease (ALL versus NHL; Fig. S4). Next, the inventors investigated whether our predictor, based on culture-associated DNAm, was also predictive of clinical outcome after adoptive T-cell transfer—although the timing of culture was relatively consistent within individual clinical trials.Interestingly, cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), which may also reflect higher graft activity, occurred significantly more frequently with CAR T cell products predicted to have a shorter culture time (p = 0.002 for CRS; p = 0.001 for ICANS; Fig. 4d). A multivariate Cox proportional hazards regression model, adjusting for clinical trial and disease, was used to test the association between predicted culture time and mortality in cancer patients (GSE179414). The estimated hazard ratios (HR) with two-sided 95% confidence intervals showed no association between predicted time in culture and overall survival (OS; HR = 0.99; CI = 0.8-1.2; p > 0.05).An individual analysis of the 35 CpGs comprising the elastic net predictor revealed that two of these CpGs were associated with overall survival (HR > 1; p < 0.01, not adjusted for multiple testing; Fig. 4e). Thus, there was a correlation between epigenetic predictions and ICANS and CRS, while only a subset of the culture-associated CpGs might be related to therapeutic outcome. Hypermethylation at specific culture-associated CpGs is an indicator of treatment success. Furthermore, the inventors investigated the association of DNAm at the 339 culture-time-associated CpGs with overall survival.Since clinical metadata were only available for samples from three clinical trials (GSE179414: NCT03144583, NCT02772198, and NCT03373071), the inventors randomly divided this dataset into a target identification cohort of 82 patients and a target validation cohort of 32 patients (Supplementary Table S4) with equal distributions regarding age, clinical trial, disease, clinical response, and CRS / ICANS. Multivariate Cox regression analysis in the target cohort revealed that 14 of the 339 CpG sites associated with culture time were associated with a higher patient mortality rate (HR > 1, p < 0.01; Fig. 5a). It is worth noting that after correction for multiple testing (for 339 CpGs), none of the individual CpGs reached statistical significance (Supplementary Table S2). However, five of the 14 CpGs were also associated with overall survival in the target validation cohort, despite the relatively small sample size (Fig.5b), and three of these CpGs were associated with the genes NR3C2, TCF7, and CASP10. The effect of DNAm on overall survival was also evident in Kaplan-Meier survival plots, in which samples were divided into quartiles according to DNAm values ​​(Fig. S5). To derive an epigenetic predictor of therapeutic outcome, we subsequently trained a model based on the 14 culture-time-associated CpGs that might be related to overall survival. This model was trained on the initial training dataset of T cells and CAR-T cells during culture expansion (multivariable rigid linear regression model with 10-fold cross-validation; Supplementary Table S2). This model again showed a high correlation between the predicted and actual time in culture in the training dataset (r2 = 0.94) and in the independent predictor validation set (r2 = 0.70; Fig. 5c).The inventors then tested this rigid 14-CpG model for an association with overall survival in the target identification clinical dataset (n = 82; HR = 1.37; p < 0.0001) and in the target validation cohort (n = 32; HR = 1.34; p < 0.01; Fig. 5d). The clear association of this epigenetic signature for culture expansion with OS was also evident in the Kaplan-Meier plots (Fig. 5e). The time of CAR T cell expansion was relatively consistent across all three clinical trials (NCT03144583, NCT02772198, and NCT03373071), thus reflecting not only culture time but also sample-inherent differences. As expected, the predicted time in culture varied between studies depending on the different production regimens (Fig. S5a), while overall survival time did not differ significantly between clinical studies (Fig. S5b).Notably, epigenetic predictions were also associated with overall survival within individual clinical trials, despite the relatively small sample sizes (Fig. S5c). Furthermore, CAR T cell products from patients experiencing CRS or ICANS after treatment were predicted to have lower overall culture expansion (p < 0.01 in the target identification cohort; Fig. 5f). Thus, culture time-associated DNAm changes in CAR T cell transplants are related to post-infusion clinical outcomes. CAR T cell manufacturing is a complex, multi-step process that should be further optimized to increase therapeutic success.The study described here shows that in vitro expansion of T cells induces DNAm alterations at specific sites in the genome, and this appears to be associated with reduced long-term survival after administration of CAR T cell products. Another recent study has also identified epigenetic signatures of prognostic relevance that have been linked to CAR vector transduction (8)—and are therefore unrelated to the culture time-related DNAm alterations we investigated. Epigenetic biomarkers are therefore powerful tools to further optimize the CAR T cell manufacturing process and gain insights into the molecular mechanisms relevant for therapeutic success.The inventors found that most culture-associated DNAm changes were hypermethylations—particularly at CpGs, which continuously changed during culture expansion. Remarkably, very similar DNAm changes were observed during T cell and CAR T cell expansion, with different culture media, and during small-scale manual production or large-scale automated production. If this hypermethylation is relevant to the reduced therapeutic success, it might be beneficial to block de novo DNAm during CAR T cell expansion. Indeed, it was recently shown that low-dose priming with decitabine—an inhibitor of DNA methyltransferases (DNMTs)—reduced fatigue, maintained the memory phenotype and effector functions, and enhanced the antitumor response in mice (31, 32).Furthermore, it has been demonstrated in mice that de novo DNAm through Dnmt3a controls the fate decisions of early effector CD8+ T cells after activation (33, 34). We recently demonstrated that knockout of DNMT3A abolishes almost all de novo DNAm during hematopoietic differentiation of induced pluripotent stem cells (iPSCs) (35). Another study recently showed that deletion of DNMT3A in human CAR-T cells prevents exhaustion and enhances anti-tumor activity (36). Notably, these studies have also revealed marked alterations in genes similar to those described here, including TCF7 and LEF1. Therefore, the inventors assume that the knockout of DNMT3A actually blocks the culture-associated DNAm changes during CAR-T cell expansion, leading to improved CAR-T cell functionality.The same epigenetic process might even contribute to CAR T cell exhaustion after infusion. The importance of epigenetic regulation for CAR T cell function in vivo is highlighted by a patient report of clonal CAR T cell expansion with biallelic TET2 disruption (37). Efforts have been made to maintain CAR T cells in a naive or memory stem cell fate to enhance anti-tumor efficacy and long-term persistence in vivo (38, 39). A recent study examined longitudinal DNA changes in CAR T cells after infusion and found continuously increasing epigenetic changes that can be directly linked to exhaustion, including the suppression of TCF7 and LEF1 and the demethylation of TOX (40).This striking overlap supports the hypothesis that DNAm alterations continuously acquired during culture expansion persist after transfusion and contribute to CAR T cell exhaustion. Unfortunately, the term "T cell exhaustion" encompasses a heterogeneous group of mechanisms that are poorly defined (41). CAR T cell dysfunction is mainly characterized by immunophenotypic changes; however, molecular phenotyping might be more appropriate. It has been suggested that dynamic changes in chromatin accessibility (analyzed by ATAC-seq) and three-dimensional chromosome conformation precede changes in gene expression, particularly near exhaustion-associated genes (42). A notable example is the TOX gene, which has been shown to drive T cell exhaustion in mice (43).Furthermore, several genes that become hypermethylated during culture expansion are relevant for the maintenance of a naive or memory-like phenotype of T cells, including TCF7, FOXO3, and PXN (44, 45). Thus, the epigenetic changes during in vitro culture expansion could reflect CAR T cell exhaustion—or they could even be related to replicative senescence. Primary cells can only be cultured for a limited number of passages before they cease proliferation and enter a state of replicative senescence. For other cellular therapeutics, such as MSCs and HSPCs, in vitro expansion is known to lead to a continuous decline in proliferation, differentiation potential, and therapeutic potential (46, 47).Interestingly, long-term culture-associated DNAm alterations in MSCs and fibroblasts can be completely reversed by reprogramming iPSCs, and they are then gradually reacquired upon redifferentiation (15, 48). Here, the inventors show that culture-associated DNAm alterations in CAR T cells are more closely related to those in HSPCs, supporting the notion of differences in molecular alterations between hematopoietic and mesenchymal lineages. Such signatures can also be influenced by changes in cellular composition, and the inventors therefore excluded CpGs that might more likely reflect changes in the CD4 / CD8++ ratio or chronological age. However, it remains largely unclear how culture-associated DNAm alterations are regulated.When the inventors analyzed culture-associated DNAm alterations in CAR-T cells by targeted amplicon sequencing, they found that the neighboring CpGs are not coherently modified on the same DNA strand. This may be unexpected, as targeted epigenetic regulation by binding an epigenetic writer would most likely also modulate the neighboring CpGs of the differentially methylated genomic region (49). Therefore, the culture-associated DNAm alterations in CAR-T cells may rather reflect dysregulation, similar to those observed for epigenetic drifts during organismal aging (50, 51). A better understanding of how DNAm alterations develop during culture expansion could shed light on the underlying process and provide more elegant therapeutic mechanisms to generate improved therapeutic CAR-T cell products and prevent their exhaustion.The inventors have shown that the DNAm level at various culture-associated CpGs in CAR-T cells is significantly associated with overall survival after infusion. The 14 CpG signature could be a powerful predictor, not only for estimating therapeutic response but also for further fine-tuning culture conditions for a better therapeutic outcome. It is worth noting that in all clinical trials the inventors tested with the predictor, the culture time was fairly consistent—thus, there is patient-specific variation that should be further explored. The results clearly indicate that a shortened cell culture time during manufacturing could be beneficial. However, whether a very short in vitro culture of CAR-T cells actually leads to better clinical outcomes remains to be demonstrated through appropriate clinical trials.Indeed, there may be a paradigm shift toward lower culture expansion of CAR T cells. Since the FDA approval of the first CAR T cell therapy in 2017, there has been a growing realization that phenotype is more important than absolute cell count. While infusion of a higher number of effector T cells can lead to higher initial response rates, long-term in vivo persistence of stem-like CAR T cells is required to avoid relapse (52). When mice were treated with CAR T cells generated without in vitro activation and expansion, their response was superior to that of CAR T cells generated using standard protocols (53). Recently, multicenter Phase I trials were initiated to evaluate the safety and preliminary efficacy of CAR T cells produced in less than two days in patients with diffuse large B-cell lymphoma (NCT03960840) or multiple myeloma (NCT04318327).Therefore, it will be interesting to analyze whether a shortened manufacturing time improves therapeutic efficacy. Several recent studies suggest that inhibiting de novo methylation during CAR-T cell expansion could block aberrant DNAm and thereby increase therapeutic efficacy. The results of the current study provide another piece of the puzzle, demonstrating that these deleterious DNAm alterations are continuously acquired during culture expansion in the manufacturing process. The data support the notion that the design of the manufacturing process is an important aspect contributing to the failure of CAR-T cell treatment. A shortened culture time to avoid dysfunctional methylation programs could be a promising manufacturing strategy to improve the therapeutic efficacy of adoptive cell therapies.The epigenetic signature for culture-associated DNAm described here provides a biomarker for quality control of CAR T cell production and for identifying patients at risk for adverse events.

[0002] Supplementary^Tables^ Table^S1.^Predictor^for^time^in^culture^(35^CpGs,^elastic^ net). Target Gene Name Gene Group Coefficient Methyl Change (Intercept) -8.49222 cg14417920 Not classified - MED15 Body Not classified LOC100188949 TSS1500 Body Not MBP LIME1 Not TSC22D4 Table^S2.^Predictors^associated^with^survival^for^time^in^culture^(14^CpGs).^ Results were calculated from the target identification cohort (subset of GSE179414, n=82). HR = hazard ratio; CI = confidence interval; OS = overall survival; FDR = FDR-adjusted p-value derived from the Wald test. Target Gene Name Gene Methyl.- HR for Death OS OS FDR Group Coefficient Change (95%CI) p-value p-value (Intercept) -17.47 cg08364283 PXN body -1.71 hyper 449 (6 - 342010.00119 0.19 cg03898320 KLF7 TSS1500 2.97 hyper 1e5 (58 - 2e8)0.00126 0.19 1 e4 (16 0.19 cg21108925 16.56 hyper 710 (5 - 0.00256 0.19 0.19 cg14117392 Not 3.75 hyper 238 (5 - 1202) 0.00516 0.19 4098 (27 - 0.19 cg13298528 CXCR5 body 6.42 hyper 23e4 (14 - 5e7)0.00544 0.19 cg06175418 TCF7 body 31 (2 - 0.19 cg04867484 NR3C2 body -1.69 hyper 2199 (10 -0.19 cg18387515 Not classified 0.20 12067423 Not 777 hr 648 (10 - 4e4) 000831 022 cg classified -. ype . .22 cg13789303 FOXO3 body 211 hyper 4e5 (48 - 4e9)0.00946 0.22 Table S3. Primers for bisulfite amplicon sequencing. Primer Chr Pos. Start Pos. End Size Sequence TOX 8 60029898 60030234 391 Forward 5`-GGGATTTTTAATATTTGTTTGGTGG-3' 15 5`-GAATTTAATAGATGTTTTTTGAGG-3' 22 5`-TAAGTATTAGATAGTGTGTAGGAG-3' Table^S4.^Distribution^of^clinical^samples^for^training and^ validation sets. Entire cohort Test cohort for the test cohort for the patient characteristics GSE179414 Target identification Target validation (n = 114) (n = 82) (n = 32) Sex, No. (%) Male 68 (59.65) 46 (56.10) 22 (68.75) Female 46 (40.35) 36 (43.90) 10 (31.25) Median age (range), y 23.5 (3 - 70) 22.4 (3 - 70) 25.8 (4 - 69) Age, No. (%), y <18 42 (36.84) 31 (37.80) 11 (34.38) 18-29 27 (23.68) 19 (23.17) 8 (25.00) 30-59 34 (29.82) 25 (30.49) 10 (31.25) ≥60 11 (9.65) 7 (8.54) 3 (9.38) Diagnosis, No. (%) B-ALL 77 (67.54) 55 (67.07) 22 (68.75) B-NHL 37 (32.46) 27 (32.93) 10 (31.25) DLBCL 20 (17.54) 14 (17.07) 6 (18.75) PMBCL 11 (9.65) 9 (10.98) 2 (6.25) Follicular lymphoma 4 (3.51) 3 (3.66) 1 (3.13) Burkitt lymphoma 1 (0.88) 1 (1.22) 0 (0.00) Mantle cell lymphoma 1 (0.88) 0 (0.00) 1 (3.13) Response, number (%) Complete response 74 (64.91) 53 (64.63) 21 (65.63) Partial response 16 (14.04) 11 (13.41) 5 (15.63) Stable disease 9 (7.89) 6 (7.32) 3 (9.38) Course of disease 15 (13.16) 12 (14.63) 3 (9.38) CRS, No. (%) Class 0 41 (35.96) 29 (35.37) 12 (37.50) Class 1 46 (40.35) 34 (41.46) 12 (37.50) Class 2 13 (11.40) 9 (10.98) 4 (12.50) Class 3 8 (7.02) 6 (7.32) 2 (6.25) Class 4 4 (3.51) 3 (3.66) 1 (3.13) Class 5 2 (1.75) 1 (1.22) 1 (3.13) ICANS, No. (%) Class 0 87 (76.32) 65 (79.27) 22 (68.75) Class 1 11 (9.65) 7 (8.54) 4 (12.50) Class 2 5 (4.39) 3 (3.66) 2 (6.25) Class 3 6 (5.26) 2 (2.44) 4 (12.50) Class 4 5 (4.39) 5 (6.10) 0 (0.00) Class 5 0 (0.00) 0 (0.00) 0 (0.00) Origin of CAR-T cells NCT02772198 43 (37.72) 31 (37.80) 12 (37.50) NCT03144583 45 (39.47) 32 (39.02) 13 (40.63) NCT03373071 26 (22.81) 19 (23.17) 7 (21.88) References 1. Elsallab M, Levine BL, Wayne AS, Abou-El-Enein M. Pre- and post-approval performance of CAR T-cell products in hematological malignancies. Lancet Oncol.2020;21(2):e104-e16. 2. Kersten MJ, Spanjaart AM, Thieblemont C.CD19-gerichtete CAR-T-Zell- Therapie bei B-Zell-NHL. Curr Opin Oncol.2020;32(5):408-17. 3. Maude SL, Laetsch TW, Buechner J, Rives S, Boyer M, Bittencourt H, et al. Tisagenlecleucel in Children and Young Adults with B-cell Lymphoblastic Leukemia. N Engl J Med.2018;378(5):439-48. 4 Ferrer G, Álvarez-Errico D, Esteller M. Biological and Molecular Factors Predicting Response to Adoptive Cell Therapies in Cancer. JNCI: Journal of the National Cancer Institute.2022;114(7):930-9. 5. Greenberg MVC, Bourc'his D. The diverse roles of DNA methylation in mammalian development and disease. Nat Rev Mol Cell Biol. 2019;20(10):590-607. 6. Mattei AL, Bailly N, Meissner A. DNA-Methylierung: eine historische Perspektive. Trends Genet.2022;38(7):676-707. 7. Wagner W. How to Translate DNA Methylation Biomarkers Into Clinical Practice. Front Cell Dev Biol.2022;10:854797. 8. Garcia-Prieto CA, Villanueva L, Bueno-Costa A, Davalos V, Gonzalez- Navarro EA, Juan M, et al.Epigenetic Profiling and Response to CD19 Chimeric Antigen Receptor T-Cell Therapy in B-cell Malignancies. J Natl Cancer Inst. 2022; 114 (3): 436–45. 9. Franzen J, Georgomanolis T, Selich A, Kuo CC, Stoger R, Brant L, et al. DNA methylation changes during long-term in vitro cell culture are caused by epigenetic drift. Commun Biol. 2021; 4 (1): 598. 10. Koch CM, Joussen S, Schellenberg A, Lin Q, Zenke M, Wagner W. Monitoring of Cellular Senescence by DNA Methylation at Specific CpG sites. Aging Cell. 2012; 11 (2): 366–9. 11. Han Y, Franzen J, Stiehl T, Gobs M, Kuo CC, Nikolic M, et al. New targeted approaches for epigenetic age predictions. BMC Biol. 2020;18(1):71. 12. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30(10):1363–9. 13. Weidner CI, Walenda T, Lin Q, Wölfler MM, Denecke B, Costa IG, et al.Hematopoietic Stem and Progenitor Cells Acquire Distinct DNA- Hypermethylation During in vitro Culture. Scientific Reports.2013;3:3372. 14 Franzen J, Zirkel A, Blake J, Rath B, Benes V, Papantonis A, et al. Senescence-associated DNA methylation is stochastically acquired in subpopulations of mesenchymal stem cells. Aging Cell.2017;16(1):183. 15. Koch CM, Reck K, Shao K, Lin Q, Joussen S, Ziegler P, et al. Pluripotent stem cells escape from senescence-associated DNA methylation changes. Genome Res.2013;23(2):248-59. 16. Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw.2010;33(1):1-22. 17. Gaidatzis D, Lerch A, Hahne F, Stadler MB. QuasR: Quantifizierung und Annotation von kurzen Reads in R. Bioinformatics.2015;31(7):1130-2. 18. Kim D, Langmead B, Salzberg SL. HISAT: ein schneller Spleiß-Aligner mit geringem Speicherbedarf. Nat Methods.2015;12(4):357-60. 19. Robinson MD, McCarthy DJ, Smyth GK.edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics.2010;26(1):139-40. 20. Law CW, Chen Y, Shi W, Smyth GK. voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol.2014;15. 21. Krueger F, Andrews SR. Bismark: ein flexibler Aligner und Methylierungs- Caller für Bisulfite-Seq-Anwendungen. Bioinformatics.2011;27(11):1571-2. 22. Marks-Konczalik J, Dubois S, Losi JM, Sabzevari H, Yamada N, Feigenbaum L, et al. IL-2-induced activation-induced cell death is inhibited in IL-15 transgenic mice. PNAS.2000;97. 23. Wagner JR, Busche S, Ge B, Kwan T, Pastinen T, Blanchette M. The relationship between DNA methylation, genetic and expression inter- individual variation in untransformed human fibroblasts. Genome Biol. 2014;15(2):R37. 24. Shan Q, Li X, Chen X, Zeng Z, Zhu S, Gai K, et al.Tcf1 and Lef1 continuously monitor the identity and function of mature CD8(+) T cells by organizing the genomic architecture. Nat Commun. 2021;12(1):5863. 25. Gagnon J, Ramanathan S, Leblanc C, Cloutier A, McDonald PP, Ilangumaran S. IL-6, in synergy with IL-7 or IL-15, stimulates TCR-independent proliferation and functional differentiation of CD8+ T lymphocytes. J Immunol. 2008;180(12):7958–68. 26. Morris EC, Neelapu SS, Giavridis T, Sadelain M. Cytokine release syndrome and associated neurotoxicity in cancer immunotherapy. Nat Rev Immunol. 2022;22(2):85–96. 27. Gumber D, Wang LD. Improving CAR-T immunotherapy: Overcoming the challenges of T cell exhaustion. EBioMedicine. 2022;77:103941. 28. Wagner W. The Link Between Epigenetic Clocks for Aging and Senescence. Front Genet. 2019;10:303. 29. Bork S, Pfister S, Witt H, Horn P, Korn B, Ho AD, et al.DNA Methylation Pattern Changes upon Long-Term Culture and Aging of Human Mesenchymal Stromal Cells. Aging Cell.2010;9(1):54-63. 30. Gross AM, Jaeger PA, Kreisberg JF, Licon K, Jepsen KL, Khosroheidari M, et al. Methylome-wide Analysis of Chronic HIV Infection Reveals Five-Year Increase in Biological Age and Epigenetic Targeting of HLA. Mol Cell. 2016;62(2):157-68. 31 You L, Han Q, Zhu L, Zhu Y, Bao C, Yang C, et al. Decitabine-Mediated Epigenetic Reprograming Enhances Anti-leukemia Efficacy of CD123- Targeted Chimeric Antigen Receptor T-Cells. Front Immunol.2020;11:1787. 32. Wang Y, Tong C, Dai H, Wu Z, Han X, Guo Y, et al. Low-dose decitabine priming endows CAR T cells with enhanced and persistent antitumour potential via epigenetic reprogramming. Nat Commun.2021;12(1):409. 33. Ladle BH, Li KP, Phillips MJ, Pucsek AB, Haile A, Powell JD, et al. De novo DNA methylation by DNA methyltransferase 3a controls early effector CD8+ T-cell fate decisions following activation.Proc Natl Acad Sci U S A. 2016;113(38):10631-6. 34. Ghoneim HE, Fan Y, Moustaki A, Abdelsamed HA, Dash P, Dogra P, et al. De Novo Epigenetic Programs Inhibit PD-1 Blockade-Mediated T Cell Rejuvenation. Cell.2017;170(1):142-57 e19. 35. Cypris O, Franzen J, Frobel J, Gluck P, Kuo CC, Schmitz S, et al. Hematopoietic differentiation persists in human iPSCs defective in de novo DNA methylation. BMC Biol.2022;20(1):141. 36 Prinzing B, Zebley CC, Petersen CT, Fan Y, Anido AA, Yi Z, et al. Deleting DNMT3A in CAR-T-Zellen verhindert Erschöpfung und erhöht die Antitumoraktivität. Wissenschaft Translationale Medizin.2021;13. 37. Fraietta JA, Nobles CL, Sammons MA, Lundh S, Carty SA, Reich TJ, et al. Disruption of TET2 promotes the therapeutic efficacy of CD19-targeted T cells. Nature.2018;558(7709):307-12. 38. Arcangeli S, Bove C, Mezzanotte C, Camisa B, Falcone L, Manfredi F, et al.Engineering CAR T cells from naive / stem memory T lymphocytes enhances antitumor responses while limiting cytokine release syndrome. The Journal of Clinical Investigation. 2022;132(12). 39. Zebley CC, Gottschalk S, Youngblood B. Rewriting history: Epigenetic reprogramming of CD8(+) T cell differentiation to improve immunotherapy. Trends Immunol. 2020;41(8):665–75. 40. Zebley CC, Brown C, Mi T, Fan Y, Alli S, Boi S, et al. CD19 CAR T cells undergo DNA methylation programming of exhaustion in patients with acute lymphoblastic leukemia. Cell Rep. 2021;37(9):110079. 41. Beltra JC, Manne S, Abdel-Hakeem MS, Kurachi M, Giles JR, Chen Z, et al. Developmental Relationships of Four Exhausted CD8(+) T Cell Subsets Reveals Underlying Transcriptional and Epigenetic Landscape Control Mechanisms. Immunity.2020;52(5):825-41 e8. 42. Gennert DG, Lynn RC, Granja JM, Weber EW, Mumbach MR, Zhao Y, et al.Dynamic chromatin regulatory landscape of human CAR T cell exhaustion. Proc Natl Acad Sci US A.2021;118(30). [ PubMed ] 43. Scott AC, Dundar F, Zumbo P, Chandran SS, Klebanoff CA, Shakiba M, et al. TOX is a critical regulator of tumor-specific T-Cell differentiation. Nature.2019;571(7764):270–4. 44. Sullivan JA, Kim EH, Plisch EH, Peng SL, Suresh M. FOXO3 regulates CD8 T-cell handling through T-cell intrinsic mechanisms. PLoS Pathology. 2012;8(2):e1002533. [ PubMed ] 45. Gauthier L, Corgnac S, Boutet M, Gros G, Validire P, Bismuth G, et al. Paxillin Binding to the Cytoplasmic Domain of CD103 Promotes Cell Adhesion and Effector Functions for CD8(+) Resident Memory T Cells in Tumors. Cancer Res.2017;77(24):7072-82. [ PMC free article ] [ PubMed ] 46. Campisi J, d'Adda di Fagagna F. Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol.2007;8(9):729–40. [ PubMed ] 47. Fernandez-Rebollo E, Franzen J, Goetzke R, Hollmann J, Ostrowska A, Oliverio M, et al.Senescence-Associated Metabolomic Phenotype in Primary and iPSC-Derived Mesenchymal Stromal Cells. Stem cell reports. 2020;14(2):201-9. 48. Frobel J, Hemeda H, Lenz M, Abagnale G, Joussen S, Denecke B, et al. Epigenetic rejuvenation of mesenchymal stromal cells derived from induced pluripotent stem cells. Stem Cell Reports.2014;3(3):414-22. 49. Hervouet E, Peixoto P, Delage-Mourroux R, Boyer-Guittaut M, Cartron PF. Spezifische oder unspezifische Rekrutierung von DNMTs für die DNA- Methylierung, ein epigenetisches Dilemma. Clin Epigenetics.2018;10:17. 50. Teschendorff AE, West J, Beck S. Age-associated epigenetic drift: implications, and a case of epigenetic thrift? Hum Mol Genet.2013;22(R1):7- 15. 51. Hernando-Herraez I, Evano B, Stubbs T, Commere PH, Jan Bonder M, Clark S, et al. Ageing affects DNA methylation drift and transcriptional cell-to-cell variability in mouse muscle stem cells. Nat Commun.2019;10(1):4361. 52.Biasco L, Izotova N, Rivat C, Ghorashian S, Richardson R, Guvenel A, et al. Clonal expansion of T-memory stem cells determines early antileukemic responses and long-term CAR-T cell persistence in patients. Nat Cancer.2021;2(6):629-42. 53. Ghassemi S, Durgin JS, Nunez-Cruz S, Patel J, Leferovich J, Pinzone M, et al. Rapid manufacturing of non-activated potent CAR T cells. Nat Biomed Eng.2022;6(2):118-28. 54. Salas LA, Koestler DC, Butler RA, Hansen HM, Wiencke JK, Kelsey CT, et al. An optimized library for reference-based deconvolution of whole-blood biospecimens assayed using the Illumina HumanMethylationEPIC BeadArray. Genome Biol.2018;19(1):64. 55. Arpon A, Milagro FI, Ramos-Lopez O, Mansego ML, Santos JL, Riezu-Boj JI, et al. Epigenome-wide association study in peripheral white blood cells involving insulin resistance. Sci Rep.2019;9(1):2445. 56. van den Berg JH, Heemskerk B, van Rooij N, et al.Tumor infiltrating lymphocytes (TIL) therapy in metastatic melanoma: boosting of neoantigen- specific T cell reactivity and long-term follow-up. Journal for ImmunoTherapy of Cancer 2020;8:e000848. doi: 10.1136 / jitc-2020-000848. Ergänzende Referenzen: 1. Houseman, E.A., Molitor, J. & Marsit, C.J. Reference-free cell mixture adjustments in analysis of DNA methylation data. Bioinformatics 30, 1431- 1439 (2014). 2. Salas, L.A. et al. An optimized library for reference-based deconvolution of whole-blood biospecimens assayed using the Illumina HumanMethylationEPIC BeadArray. Genome Biol 19, 64 (2018). 3. McCartney, D.L. et al. Identification of polymorphic and off-target probe binding sites on the Illumina Infinium MethylationEPIC BeadChip. Genom Data 9, 22-4 (2016). 4. Pidsley, R. et al. Critical evaluation of the Illumina MethylationEPIC BeadChip microarray for whole-genome DNA methylation profiling. Genome Biol 17, 208 (2016).

Claims

RWTH Aachen 23513WO Patent claims 1. Method for controlling the quality of T cells during and / or after cultivation of these T cells, and / or for controlling and optimizing the cultivation of T cells, the method comprising: a) isolating at least one nucleic acid molecule of at least one T cell during and / or after cultivation; b) determining the degree of methylation of at least one CpG dinucleotide of the nucleic acid molecule, wherein the CpG dinucleotide is selected from the group consisting of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824, cg13789303 and at least one CpG dinucleotide located within a region of 500 nucleotides upstream and / or downstream of each of said CpG dinucleotides;c) Comparing the methylation level determined under b) with at least one reference value corresponding to the methylation level of the CpG dinucleotide of uncultured primary T cells, the result of the comparison directly indicating whether the cultured T cells are suitable for therapeutic purposes or whether the cultivation conditions should be modified.

2. The method according to claim 1, characterized in that the T cells are recombinant T cells comprising at least one gene encoding a modified antigen receptor, wherein the gene preferably encodes an adapted T cell receptor (TCR) or a chimeric antigen receptor (CAR).

3. The method according to claim 1 or 2, characterized in that at least one CpG dinucleotide is located within a region of 400, 300, 200 or 100 nucleotides upstream and / or downstream of each of the CpG dinucleotides cg08364283, cg03898320, cg20606093,; 2 cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824 and cg13789303.

4. The method according to claim 1, 2 or 3, characterized in that suitability for therapeutic purposes is assumed if the methylation level determined under b) substantially corresponds to the reference value, or suitability for therapeutic purposes is assumed not to be present if the methylation level determined under b) is increased compared to the reference value.

5. The method according to claim 1, 2, or 3, characterized in that the cultivation conditions can remain unchanged if the degree of methylation determined under b) essentially corresponds to the reference value, or the cultivation conditions should be changed if the degree of methylation determined under b) is increased compared to the reference value. 6.Method according to claim 5, characterized in that the cultivation conditions are changed such that the changed conditions comprise a change in the composition of the culture medium.

7. Method according to claim 5 or 6, characterized in that the cultivation conditions are changed such that the changed conditions comprise a shortening of the cultivation time.

8. Method according to claim 5, 6 or 7, characterized in that the cultivation conditions are changed such that the methylation of the CpG dinucleotide is inhibited.

9. Use of at least one nucleic acid molecule which comprises at least one of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528. 3 cg06175418, cg04867484, cg18387515, cg12067423, cg09801824 and cg13789303 and / or at least one CpG dinucleotide located within a region of 500 nucleotides upstream and / or downstream of each of said CpG dinucleotides, for controlling the quality of T cells during and / or after cultivation of these T cells.

10. Use of at least one nucleic acid molecule which comprises at least one of the CpG dinucleotides cg08364283, cg03898320, cg20606093, cg21108925, cg07279377, cg14117392, cg04455867, cg13298528, cg06175418, cg04867484, cg18387515, cg12067423, cg09801824 and cg13789303 and / or at least one CpG dinucleotide which is located within a region of 500 nucleotides upstream and / or downstream of each of the said CpG dinucleotides, for controlling and optimising a method for Cultivation of T cells.

Images