In animal models, elimination of the senescent cells in the hematopoietic stem cells (HSCs) compartment leads to the rejuvenation of hematopoiesis. Whether this treatment principle can be applied to the human system remains controversial. The identification of senescent cells in human bone marrow poses another major challenge. To address these questions, we have studied hematopoietic stem and progenitor cells (HSPCs, CD34+) from the bone marrow of 15 healthy human subjects (age range: 19-74 years). Single-cell RNA sequencing, functional transcriptome analysis, and development trajectory studies were performed. In a previous report, we demonstrated the accumulation of a senescent population in the aging HSC compartment. The present study focuses on the differences with age downstream in the lymphoid trajectory. While a reduction in B progenitors in the early lymphoid compartment can be confirmed, the accumulation of a lymphoid cluster downstream upon aging is novel and remarkable. This cluster comprises cells with a significant deficiency in B differentiation markers, as well as 9.4% cells with transcriptome signatures of memory-like natural killer (NK) progenitors. Applying our analysis algorithm to other human bone marrow datasets from the literature, we are able to validate the presence of this unique cluster in aged lymphoid progenitors. The accumulation of a population comprising cells defective in B differentiation potential, as well as cells with transcriptome features of memory-like NK progenitors represents a novel hallmark for senescence in the late development trajectory of human lymphoid compartment.

To identify the differences between aged and young human hematopoiesis, we performed a direct comparison of aged and young human hematopoietic stem and progenitor cells (HSPCs). Alterations in transcriptome profiles upon aging between humans and mice were then compared. Human specimens consist of CD34+ cells from bone marrow, and mouse specimens of hematopoietic stem cells (HSCs; Lin− Kit+ Sca1+ CD150+). Single-cell transcriptomic studies, functional clustering, and developmental trajectory analyses were performed. A significant increase in multipotent progenitor 2A (MPP2A) cluster is found in the early HSC trajectory in old human subjects. This cluster is enriched in senescence signatures (increased telomere attrition, DNA damage, activation of P53 pathway). In mouse models, the accumulation of an analogous subset was confirmed in the aged LT-HSC population. Elimination of this subset has been shown to rejuvenate hematopoiesis in mice. A significant activation of the P53–P21WAF1/CIP1 pathway was found in the MPP2A population in humans. In contrast, the senescent HSCs in mice are characterized by activation of the p16Ink4a pathway. Aging in the human HSC compartment is mainly caused by the clonal evolution and accumulation of a senescent cell cluster. A population with a similar senescence signature in the aged LT-HSCs was confirmed in the murine aging model. Clearance of this senescent population with senotherapy in humans is feasible and potentially beneficial.

Background: In mouse model for aging studies, clearance of senescent cells in the hematopoietic stem cell (HSC) compartment has been shown to rejuvenate the hematopoietic system. To identify the contribution of senescent cells to aging of human hematopoiesis, we have performed a direct comparison of aged and young human hematopoietic stem and progenitor cells (HSPCs). In a subsequent step, we compared the alterations in transcriptome profiles upon aging between humans and mice to determine the mechanistic differences in the aging process between the two species. Methods: Human specimens: HSPCs (CD34+) were isolated from the bone marrow of healthy human subjects (old = 59 to 74 years, n=7; young = 21 to 31 years, n=5). Mousespecimens: Bone marrow cells were harvested from 10 young (6 to 8 weeks old) and 10 aged (100 weeks old) mice. Murine hematopoietic stem cells (HSCs) were isolated via flow cytometry sorting (characterized as Lin- Kit+ Sca1+ CD150+ cells). Single cell transcriptome studies of these specimens were performed, followed by functional clustering and developmental trajectory analyses. Results: We have identified two novel subsets of HSPCs in aged human subjects. A cluster, designated as multipotent progenitor 2A (MPP2A), is found in the early phase of HSC aging in old subjects, and another cluster “PreB_X” downstream in the lymphoid trajectory. The MPP2A cluster is highly enriched in senescence signature, as demonstrated by their transcriptome profile showing increased telomere attrition, DNA damage, activation of P53 pathway and cell cycle arrest, with significant elevations of the following genes: ATF3, CDKN1A, CLU, FOSB, ID1, JUN, MAFF, PLK3. In mouse model, the accumulation of an analogous subset with senescence signature (MPP2A analogs) was confirmed in the aged LT-HSC population, with significantly increased expressions of Atf3, Clu, Fosb, Id1, Jun, Plk3. Elimination of this subset has been shown to rejuvenate hematopoiesis in the same animal model. Remarkable is a consistent and significant upregulation of P53 pathway, formation of senescence-associated heterochromatin foci (SAHF), and increased expression of CDKN1A in aged human HSC compartment, and specifically in the MPP2A cluster. All these results indicate an activation of the P53-P21WAF1/CIP1 pathway as a main mechanism for cell cycle arrest in human senescent HSCs. This is in contrast to the findings in mouse model, where increased levels of Cdkn2a, Cdkn2b, and activation of the p16INK4a pathway represent the hallmarks of senescence. Applying our analysis algorithm and parameters, we were able to validate an accumulation of MPP2A analogous subset upon aging in other human HSC datasets published in the literature. Conclusion: Aging in human HSC compartment is mainly caused by the clonal evolution and accumulation of a senescent cell cluster MPP2A. The latter is characterized by increased telomere attrition, cell cycle arrest, and a remarkable up-regulation of CDKN1A. The identification of a senescent subset in human HSC compartment, in analogy to a similar population in LT-HSC in mice indicate that clearance of this senescent population with senotherapy is feasible and potentially beneficial. For humans, however, the P53-P21CIP1 positive cells, and not the p16INK4a positive cells as in mice, represent the main target.

(A) IFN-𛾠ELISpot raw data of all HLA-A*02:01 patients analyzed (n = 7) showing the IFN-𛾠spots / 1x105 T-cells for patient cells against TAAs vs. patient cells against the negative control (human immunodeficiency virus (HIV) gag/pol, 9 amino acids). Grey bars indicate a significantly increased immunogenicity compared to negative control (solid line). (B) Homogenous levels of stimulation while comparing the total immune response by mean of IFN-𛾠spot numbers among astrocytomas (red) and oligodendrogliomas (blue) to a certain reactive epitope. (n.s., not significant, *, p < 0.05; **, p < 0.01)

(A) Significantly increased immune responses in IDHmut lower-grade glioma patients (P) vs. healthy donors (HD) seen for the potential T-cell target antigens CRKII, CFL1, CNTN1, NME2, and TKT. Homogenous levels of stimulation while comparing IFN-𛾠spot numbers among astrocytomas (red) and oligodendrogliomas (blue) to a certain TAA as well as (B) the total immune response. (C) IFN-𛾠ELISpot raw data showing the spot count / 1x105 T-cells upon stimulation with autologous dendritic cells loaded with antigens (CRKII, CFL1, CNTN1, NME2, or TKT, black bar) vs. negative control (IgG1, grey bar) for each patient (P) (n = 16). Highlighted in green: significant immune responses (p < 0.05) with an > 2-fold increased spot count compared to the negative control (IgG) (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

Adhesion of hematopoietic stem and progenitor cells (HSPCs) to the bone marrow niche plays critical roles in the maintenance of the most primitive HSPCs. The interactions of HSPC-niche interactions are clinically relevant in acute myeloid leukemia (AML), because (i) leukemia-initiating cells adhered to the marrow niche are protected from the cytotoxic effect by chemotherapy and (ii) mobilization of HSPCs from healthy donors' bone marrow is crucial for the effective stem cell transplantation. However, although many clinical agents have been developed for the HSPC mobilization, the effects caused by the extrinsic molecular cues were traditionally evaluated based on phenomenological observations. This review highlights the recent interdisciplinary challenges of hematolo-gists, biophysicists and cell biologists towards the design of defined in vitro niche models and the development of physical biomarkers for quantitative indexing of differential effects of clinical agents on human HSPCs.

Representative images of immunohistochemical stainings of the isotype controls IgG and IgG1 in normal brain (NB) tissues, astrocytomas (WHO{degree sign}II: n = 10; WHO{degree sign}III: n = 10), and oligodendrogliomas (WHO{degree sign}II: n = 10; WHO{degree sign}III: n = 10). Scale bar: 50 µm.

Supplementary Figure Legends

Peptide information of potential T cell target antigens, tested by IFN-γ ELISpot

(A) Applied gating strategy (one representative patient, NCH645) for the expression analysis of tumor-associated antigens in IDHmut lower-grade glioma and secondary glioblastoma GSCs, analyzed by flow cytometry in regard to an adequate isotype control. (B) Fluorescence intensity measurements for the isotype control and the antigen expressing (positive) GSC population, showing counts in percentage [%] over the mean PE-intensity.(C) Gating strategy, including the markers pacific orange (PO), CD3, and CD8, used for quantification of antigen-specific peripheral cytotoxic CD3+CD8+ T-cells by flow cytometry and (D) antigen-loaded pMHC I-tetramers (PE-conjugated) and the respective positive (Influenza M1) as well as negative control HLA-A*02-tetramers. GSCs: glioma stem-like cells, HIV: human immunodeficiency virus, HLA: human leukocyte antigen, MHC: major histocompatibility complex, PE: Phycoerythrin, pos: positive, US: unstained.

(A) Semi-quantitative evaluation on immunohistochemically stained sections of the frequency of tumor cells expressing the potential T-cell target antigens CRKII, CFL1, NME2, and CNTN1 in astrocytoma (WHO{degree sign}II: n = 10, WHO{degree sign}III: n = 10) and (B) oligodendroglioma (WHO{degree sign}II: n = 10, WHO{degree sign}III: n = 10) tumor tissues.

Gating strategy to identify cytokine-secreting cells by cytokine capture assay. The acquired cells were gated based on their size and granularity as lymphocytes excluding cell debris. Then we gated on single cells to remove cell duplexes and clusters. Apoptotic cells were gated out next by gating on pacific orange (PO)-negative cells (live cells). The respective T-cell (TC) population was identified by surface staining with anti-CD4 PerCP-Cy5.5, and anti-CD8-V450 antibodies. The cytokine-positive cell gate was placed according to a FMO control (exemplarily shown for IFN-ð›¾) and the frequency of cytokine-secreted cells in samples stimulated with IgG1 peptide (negative control) or the antigen of interest (exemplarily shown for CFL1) were determined based on that gate.

Supplementary Figures S1-S1-; Supplementary Tables S1-S2.

Background: Acute myeloid leukemia (AML) is a hematopoietic malignancy characterized by uncontrolled proliferation and impaired differentiation of myeloid progenitor cells. With a 5-year overall survival of approx. 30%, the prognosis of patients with acute myeloid leukemia remains poor, in particular for those with relapse or refractory disease. The adoptive transfer of T-cells genetically modified to express a chimeric antigen receptor (CAR) have shown impressive remission rates in patients with B-cell lymphoid malignancies. In contrast, CAR T-cell therapy for AML has induced durable remission only in a minority of patients so far, highlighting the need for strategies to improve anti-tumor efficacy and persistence of AML-specific CAR T-cell products. Adapting an established high throughput screening platform, we aimed to identify compounds with the potential to enhance CAR T-cell mediated cytotoxicity through interaction with tumor and/or CAR T-cells. Methods: We established a high throughput platform to determine the cytotoxicity of four different AML-specific CAR T-cell products targeting CD33 (CD33.CD28.41BB.zeta), CD70 (CD27.zeta), CLL1 (CLL1.CD28.zeta) and EMR2 (also known as ADGRE2; EMR2.41BB.zeta) in the presence of 32 compounds at five different concentrations in a 10,000-fold concentration range. CAR T-cells were generated by transduction of activated T-cells derived from peripheral blood mononuclear cells (PBMCs) of three healthy donors with a gamma retroviral vector containing the different CAR constructs. On day 11 after activation, CAR T-cells and non-CAR-transduced control T-cells were incubated with luciferase-expressing target cells at an effector-target ratio of 1:4 in 384-well plates. Viability of target cells cultured with CAR T-cells, control T-cells or in the absence of T-cells was calculated after 24 and 48 hours using a luciferase signal intensity assay. Combinatorial toxicities were calculated with the Bliss Independence model for synergistic effect and the highest single agent model for additive combination effect. Results: In quality analysis, the intensity of luminescence signals in control wells (tumor only, non-CAR-transduced control T-cells) of different plates was similar and principal component analysis (PCA) of samples derived from different days (batches) did not show significant differences ruling out relevant edge or batch effects that could negatively impact the screening results. Of all tested compounds birinapant, a peptidomimetic of second mitochondrial-derived activator of caspases (SMAC) and inhibitor of apoptosis protein (IAP) family proteins and the hypomethylating agent (HMA) decitabine showed the strongest improvement of CAR T cell cytotoxicity in an additive but also in a synergistic manner. These effects were evident in CAR T-cell products derived from all different T cell donors without significant differences. Depending on the structural composition of the CAR construct, the strongest drug-induced improvement of CAR T-cell cytotoxicity was seen either after 24 or 48 hours. The third generation CAR T- cell construct (CD33.CD28.41BB.zeta) showed enhanced cytotoxicity already after 24 hours, whereas significant anti-leukemic efficacy of CD27.zeta, CLL1.CD28.zeta and EMR2.41BB.zeta CAR T-cells became evident at the later 48 hours timepoint. Importantly, co-treatment with birinapant and decitabine particularly enhanced the cytotoxicity of EMR2-specific CAR T-cells which exhibited the lowest baseline anti-tumor efficacy compared to all other CAR constructs highlighting the potential of distinct chemical compounds to improve the functionality of AML-specific CAR T-cells. Conclusion: Our data suggest that high throughput drug interaction screens are a reliable approach to investigate the impact of chemical compounds on the functionality of AML-specific CAR T-cell products. HMA and SMAC mimetics are promising candidates that may potentially enhance the cytotoxicity of CAR T-cells. Ongoing studies are evaluating the effect of HMAs and SMAC mimetics on CAR T-cell proliferation and anti-tumor efficacy against AML cell lines and primary AML patient samples upon serial antigen stimulation. Furthermore, in-depth proteogenomic characterization to identify molecular signatures of enhanced functionality are currently being performed.

Representative multicolor stainings of CFL1, CRKII, NME2, and CNTN1 on acetone-fixed cryosections with markers for common cell types in IDHmut LGGs: anti-GFAP (tumor cells, #Z0334, DAKO), anti-CD68 (microglia/ macrophages, #M0718, DAKO), anti-CD31 (endothelial cells, #223609, BD Pharmingen). Detection was performed by using fluorochrom-conjugated secondary antibodies (anti-mouse AF647 (#A-21463, Invitrogen), anti-rabbit AF555 (#A-21428, Invitrogen), DAPI (#D1306, ThermoFisher), and the Zenon AF488 mouse IgG1 Labeling Kit (#Z25001, ThermoFisher)) according to the manufacturer`s protocol (scale bar: 50 µM, scale bar zoom: 10 µM).

Graphical representation of IFN-𛾠ELISpots using peripheral blood mononuclear T cells, co-cultivated with dendritic cells, pulsed with autologous tumor lysate fractions of (A) the first (1st) PF2D dimension and corresponding (B) second (2nd) PF2D dimension of patients NCH1390, NCH612, and NCH519a. Solid line represents mean background of negative control (PBL). F: fraction, PBL: peripheral blood lymphocytes (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

Background : Hematopoietic stem and progenitor cells (HSPCs) undergo functional changes with age. An unresolved issue is whether aging is induced by a shift in intrinsic state in which all HSPCs undergo coordinated changes in functional potential, or by clonal evolution of a senescent population.Identification of senescent cells in human hematopoiesis poses another challenge. In this study, we have identified lineage-specific, senescence-associated pathways at different developmental stages of HSPCs and have developed an “aging signature HSC” for non-primed human HSCs. Methods: HSPCs (CD34+ cells) were isolated from the bone marrow of healthy human subjects (old = >60 years; young = 21 to <35 years). Single-cell RNA-sequencing (scRNAseq) studies and downstream analysis were performed using different R packages for normalization, dimension reduction, clustering, classification, pseudotime ordering, and developmental trajectory analyses. Gene Set Enrichment Analysis, GSEA (version 4.3.2., Broad Institute, Inc) was applied to identify pathways and processes that were coordinately up- or down-regulated with aging. Results: HSPCs (CD34+ cells) were derived from 13 healthy human subjects (old subjects n = 6; young subjects n = 7). There were 26,076 genes identified. The HSPCs were ordered and classified along their developmental trajectory during differentiation processes from (1) hematopoietic stem cells (HSC) to (2) myeloid-lymphoid progenitors (MLP), (3) megakaryocyte-erythrocyte progenitors (MEP), (4) granulocytic-monocytic progenitors (GMP), (5) granulocyte- (G), and (6) monocyte- (Mono) progenitors, (7) dendritic cell (DC), (8) erythroid progenitors (E), as well as various subsets of lymphocytes (Ly1 to Ly4). The results are depicted in Figure 1A. Remarkable are the increases in HSCs (p < 0.05) and in Ly4 (p < 0.001) in old subjects. The Kolmogorov-Smirnov test showed a significant delay in the developmental “pseudotimes” of HSCs (p < 0.001) and of Ly4 (p < 0.05). We then compared the differences in GSEA between old and young subjects in each of the following clusters: (A) whole HSPCs; (B) HSPCs along myeloid pathway (mye-HSPC); (C) HSC-MLP-MEP; (D) HSC. The major findings relevant to aging are summarized in Fig. 1B. Pathways that are significantly increased in the old subjects and throughout all comparisons include: “Formation of senescence associated heterochromatin foci”, “Aging signature HSC”, “DNA damage telomere stress induced senescence”, and “TNFR1 induced NFKB signaling pathway”. “Hypoxia”, “P53 pathway”, “IL2 signaling”, “TNFA signaling via NFKB” were elevated upon enrichment of the primitive HSC population, while other pathways decreased: “Senescence associated secretory phenotype (SASP)”, “DNA methylation”. “Cell cycle” and “DNA repair” are consistently more abundant in young subjects. We analyzed the 512 cells in the HSC cluster with higher resolution and were able to identify a sub-cluster “MPP2A” that is significantly increased in old subjects (4.5% of all HSPCs and 28.5% of HSCs, versus 1.9% and 12.9% respectively in young subjects; p < 0.005). The results are depicted in the highlighted circles in Fig. 1A. This population is quiescent and is highly enriched in “aging signature HSC” in old subjects (p <0.0001). Prominent examples of these signature genes within this list are shown in Fig. 1C). Further along the trajectory, a unique population of B cells “Ly4” was found predominantly in old subjects. The transcriptome profile Ly4 is similar to immature and age-associated B cells (ABCs). This cluster of lymphoid cells may represent the precursors of ABCs described thus far only in peripheral blood, providing another evidence that the aging mechanisms at various differentiation stages are different. Conclusions: In this study, we have defined specific pathways that distinctly separate aging HSCPs from young ones, and shown that the definition of “aging signature” varies with developmental stages. The accumulation a small cluster in the non-primed HSC compartment, with “aging signature HSC” and cell cycle arrest in old subjects indicates that aging of human hematopoiesis is induced by clonal evolution. The abundance of a unique population among lymphoid precursors with expression profiles similar to ABCs demonstrates that lymphoid and myeloid progenitors age differently.

Clinical data of IDHmut LGG patients