An Aged Microenvironment Increases CAR T Cell Cytotoxicity but Impairs Therapeutic Efficacy

Background: Cancer disproportionately affects the elderly, who are often less able to tolerate traditional cytotoxic therapies, and may benefit from T cell –based immunotherapies. However, studies evaluating the efficacy of T cell immunotherapy in aged mice are limited and yield inconsistent results, while clinical data are largely retrospective.

Here, we used a murine model of Chimeric Antigen Receptor (CAR) T cell therapy to investigate how aging influences efficacy, from CAR T cell production to in vivo anti-tumor activity.

We found that aging reduced CAR T cell production yields and altered their phenotype and function. Aged CAR T cells were predominantly effector memory CD4⁺ T cells, whereas young CAR T cells were primarily central memory CD8⁺ T cells. Functionally, aged CAR T cells exhibited non-specific cytotoxicity, driven by constitutive degranulation and elevated granzyme B secretion independent of CAR expression. This phenotype was induced by the aged microenvironment, as young T cells transferred into aged hosts adopted similar behavior. In vivo, young CAR T cells efficiently reduced tumor burden in young leukemia - bearing hosts , but were not effective in aged hosts, where the aged microenvironment impaired CAR T cell persistence. Conclusion:These findings indicate that aging impacts CAR T cell therapy at multiple levels, from manufacturing to therapeutic efficacy, highlighting the need to design tailored immunotherapies for elderly patients. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint

Cancer disproportionally affe cts elderly people , and aged patients are less able to tolerate the detrimental side effects of classical therapies, such as systemic chemotherapy and irradiation, which cause excessive damage to healthy tissues. Hence, aged patients could greatly benefit from more specific therapeutic approaches such as immunotherapy based on endogenous or genetically engineered T cells to directly target tumor cells. The choice of cancer immunotherapy must therefore consider age-related changes in T cell immunity1. Chimeric antigen receptor (CAR) T cells are genetically engineered T cells designed to recognize and attack tumor -specific antigens 2. These synthetic receptors consist of an extracellular antigen-binding domain, typically a single-chain variable fragment (scFv) derived from an an tibody, linked to intracellular signaling domains, including the CD3 z activation domain and one or more co -stimulatory domains such as CD28 or 4 -1BB. Upon engagement with their target antigen, CARs initiate intracellular signaling cascades that lead to T c ell activation, proliferation, and acquisition of effector functions. Activated CAR T cells eliminate target cells through the release of cytotoxic granules (e.g., perforin and granzyme B), as well as proinflammatory cytokines and chemokines3. The expression on CAR T cells of exhaustion markers such as PD-1, Tim-3, and LAG-3, which are also elevated in aged T cells, has been associated with poor clinical outcomes 4,5. However, to date, no clinical studies have directly examined the impact of T cell aging on CAR T cell immunotherapy outcomes 6. Retrospective analyses of the Tisagenlecleucel (JULIET trial)7 and Axicabtagene Ciloleucel (ZUMA-1 trial)8 found that age did not significantly affect the efficacy of anti -CD19 CAR T -cell therapies in relapsed or refractory B -cell malignancies. Both trials included patients up to 76 years of age, and reported no significant differences in (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint overall or complete response rates between those over 65 a nd younger patients. Similarly, post-authorization studies for relapsed or refractory multiple myeloma (RR MM) 9 showed comparable outcomes between younger and older patients. These findings suggest that aging does not compromise CAR T-cell efficacy; however, older patients had a higher risk of severe adverse events, including cytokine release syndrome (CRS) and neur ological toxicity. In addition, aging increases the likelihood of CAR -T product manufacturing failure 10, raising concerns that retrospective analyses may introduce selection bias by including only older patients whose T cells were fit enough to successfully complete the CAR T -cell production process6. Preclinical studies in animal models evaluating the impact of aging on immunotherapy efficacy have yielded inconsistent results. Some studies report ed enhanced cytotoxicity in CD8+ T cells from aged mice, attributed to increased secretion of perforin and granzyme B (GzB)11,12. In contrast, other studies indicated that CAR T cells derived from aged donor mice exhibit functional impairments compared to those from young donors, including lower transduction efficiency, reduced expansion, and diminished levels of key signaling molecules such as phosphorylated ERK, Akt, Stat3, and Stat5, ultimately leading to reduced cytotoxicity13. In this study, we utilized the well -established CD19 CAR T cell model 14, and demonstrated that aging negatively impacted CAR T cell production, phenotype, and function. Aged CAR T cells were predominantly comprised of effector memory CD4 + T cells, whereas young CAR T cells primarily consisted of central memory CD8 + T cells. Functionally, aged CAR T cells exhibited non -specific cytotoxicity, targeting and killing cells independent ly of CAR expression or direct contact. This excessive killing was driven by continuous degranulation and elevated GzB secretion. The aged microenviron ment played a key role in (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint inducing this phenotype, as young T cells transferred into aged hosts displayed a similar phenotype. In studies employing the ML21 leukemia model 15, young CAR T cells effectively reduced tumor burden in young leukemia-bearing hosts but were less effective in aged hosts. Within the aged microenvironment, CAR T cells exhibited reduced p ersistence and were less efficient in reducing tumor burden. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint

Mice. Young (8-12 week old) and aged (20-23 month old) C57BL/6JOlaHsd female mice were purchased from Envigo (Israel). For aging experiments, 8 months old retired breeders were maintained for additional 12-15 months. Transgenic C57Bl/6 Rosa26tdTomato/+ OTII mice were a gift from Prof. Ziv Shulman (The Weizmann Institute of Science). All mice were housed under specific pathogen-free conditions at The Technion Pre–Clinical Research Authority and used in accordance with animal care guidelines of the Institutional Animal Care and Use Committee. Cell Lines. The A20 cell line was cultured in RPMI 1640 medium supplemented with 10% FBS (Gibco), 1% Penicillin/Streptomycin (Gibco), 1% HEPES buffer, and 50μM β-mercaptoethanol. HEK293T cells (Human embryonic kidney 293 cells) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% of FBS and 1% Penicillin/Streptomycin.The cultures were maintained at 37℃ with 5% CO2 and split every 3-4 days. Construction of retroviral vectors encoding CARs . Second-generation CAR constructs were prepared by cloning the variable region of the 1D3 hybridoma that specifically recognizes murine CD19 (mCD19 scFv) into the retroviral pBullet vector , including a gene encoding the GFP label (A gift from Prof. Zelig Eshchar’s lab) . The scFv was connected via the carboxyl terminus of the heavy chain variable domain VH region to a myc tag, a short peptide sequence used to facilitate the detection of cells that expressed the construct, CD8 hinge, CD8 transmembrane domain, and CD28-CD3z intracellular domains, and cloned into the retroviral pBullet vector, followed by an internal ribosome entry site (IRES) and GFP. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint Generating CAR T cells from aged and young mice. HEK293T cells were transfected with CAR retroviral vectors and Peco packaging plasmids, using the CalFectin reagent. Virus-containing supernatants were harvested after 48 hours of culture and used for transduction. Retroviral transduction of primary mouse T cells: Splenocytes were harvested from young (8- 12 weeks old) and aged (20 -22 months old) C57Bl/6 mice, and T cells were isolated by magnetic separation using EasySep™ Mouse T Cell Isolation Kit (STEMCELL Technologies, catalog # 19851) according to the manufacturer’s instructions. The isolated cells were resuspended in stimulation media ( complete RPMI medium supplemented with 1% non - essential amino acids, 1% Sodium Pyruvate Solution, and 100 U/ml recombinant murine IL -2 (Peprotech, catalog #212-12-50)), stimulated either using weak (0.5 µg/ml a-CD28; BioXcell BE0015-1 and 1 µg/ml a-CD3; BioXcell BE0001- 1) or strong (4 µg/ml a-CD28 and 1 µg/ml a- CD3) activation, and infected with a retrovirus, using retronectin (TaKara, Catalog #T100A ) according to the manufacturer’s instructions. CAR T cells were cultured at 37 °C and 5 % CO2 in cRPMI supplemented with IL-2 (100U/ml). CAR T cell sorting. GFP+ cells were sorted using a FACS ARIA–IIIU (BD Biosciences) or Bigfoot (Thermo Fisher Scientific) Cell Sorters. After sorting, cells were cultured in complete RPMI medium supplemented with recombinant murine IL-2 (100U/ml). Killing assays. Coculture: A20 cells were labeled with CellTrace Violet (CTV) (Thermo Fisher Scientific, Catalog# C34557) following the manufacturer’s protocol , and co -cultured with young or aged control T or CAR T cells at varying target-to-effector (T:E) ratios. A20 cell death was assessed using either live cell imaging or flow cytometry. Transwell: C ytokine-mediated killing was measured using the Corning® HTS Transwell®-96 Permeable Support with a 0.4 µm Pore Polycarbon ate Membrane (Corning, catalog #3381). Young and aged CAR T cells or control T cells were cultured in the upper chamber, while A20 cells were added to the lower (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint chamber at a 1:5 T:E ratio for 10 hours. After incubation, the A20 cells were collected and stained with Zombie NIR™ Fixable Viability Kit (Biolegend, catalog # 423105) dye to assess target cell death. Flow cytometry. For cell surface staining , cells were resuspended in separation buffer (PBS containing 2% FBS and 2 mM EDTA) and incubated on ice for 20 minutes with antibody mixture, washed, and analyzed by flow cytometry. For i ntracellular staining , cells were activated using PMA and ionomycin (Biolegend, catalog # 423301) in the presence of Golgi Stop (Brefeldin A; Biolegend, catalog # 420601) for 5 hours, and stained using a True-Nuclear Transcription Factor Buffer Set (Biolegend, catalog # 424401), following the manufacturer’s protocol. List of antibodies used: Antibody Supplier Catalog number PE anti-mouse CD4 BioLegend 100408 APC anti-mouse CD8a BioLegend 100712 PE anti-mouse CD69 BioLegend 104508 Brilliant Violet 510™ anti-mouse/human CD44 Biolegend 103043 Alexa Fluor® 700 anti-mouse CD62L Biolegend 104426 eFluor™ 450 IFN gamma Monoclonal Antibody (XMG1.2) ThermoFisher Scientific 48-7311-82 Brilliant Violet 510 anti-mouse CD107a (LAMP-1) Biolegend 121629 Alexa Fluor® 647 anti-human/mouse Granzyme B Recombinant Antibody Biolegend 396421 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint Live cell imaging. Live cell imaging was performed by labeling target A20 cells with Cell Trace Violet and co-culturing them with either young or aged control T cells or CAR T cells at a 1:1 effector-to-target ratio. Annexin V antibodies were added to the culture medium to mark apoptotic cells, and the percentage of CTV/Annexin V double -positive A20 cells was quantified over 24 hours using the Zeiss Celldiscoverer 7 . Analysis was performed using Zen software. ELISA. Supernatants collected from the killing experiments were used to quantify Granzyme B and IFNg secretion. The levels of these cytokines were determined using the Mouse Granzyme B DuoSet ELISA (R&D Systems, catalog #DY1865 -05) and IFNg DuoSet ELISA (R&D Systems, catalog #DY485-05) according to the manufacturer’s instructions. Adoptive T cell transfer . Young T cells were isolated from spleens of young C57Bl/6 Rosa26tdTomato/+ (8-12 week old) by magnetic isolation (StemCell) , and 5X106 cells were transferred i.v. into wild type C57Bl/6 young or aged recipients. ML21 Leukemia model and CAR T cell treatme nt. Young or aged mice were injected intraperitoneally (IP) with 200 mg/Kg Cyclophosphamide monohydrate (Sigma, Catalog # Propidium Iodide Solution Biolegend 421301 APC anti-mouse CD19 Antibody Biolegend 152409 Alexa Fluor® 647 anti-c-Myc Antibody Biolegend 626810 CellTrace™ Violet Cell Proliferation Kit (CTV) ThermoFisher Scientific C34557 Incucyte® Annexin V Orange Reagent for apoptosis Sartorius 4759 Zombie NIR™ Fixable Viability Kit Biolegend 423105 FITC anti-mouse CD8a BioLegend 100706 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint C0768-1G). After 1 day, 1X10 6 ML21 cells were administered intravenously (iv) to the mice. Drinking water was supplemented with 20 mg/ml doxycycline for the entire experiment duration. On day 10, the mice received i.v. injection of either 2X106 control or CAR T cells, and blood samples were collected via tail vein puncture every few days to assess tumor burden and CAR T cell persistency. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint

Suboptimal activation impairs the production of CAR T cells from aged donors To directly investigate how aging affects the efficacy of CAR T cell therapy, we utilized the CD19 CAR T cell model, which has demonstrated efficacy in both preclinical and clinical studies against CD19 + B cell malignancies 16,17. A second-generation CAR construct targeting murine CD19 (mCD19; supplemental figure 1A) was generated and cloned into the GFP - labeled retroviral pBullet vector. A Myc tag18 was included to facilitate the detection of CAR- expressing T cells. CAR T cells were prepared using standard protocols19 (figure 1A), and their functionality was validated by assessing their ability to kill A20 B cell lymphoma cells expressing CD19, demonstrating a specific, dose-dependent cytotoxicity (supplemental figure 1B). Transduction efficiency in T cells from aged mice (20 -23 months old) was below 10% and significantly lower than that in T cells from young mice (8-12 weeks old), which exhibited over 40% efficiency, as measured by GFP expression (figures 1B, 1C). To enable functional studies, we sought to enhance CAR T cell production from aged T cells. Since successful transduction requires pre-activation, standard CAR T cell protocols recommend low concentrations of anti- CD3 and anti-CD28 antibodies to minimize T cell exhaustion20. Under such "weak activation" conditions (0.5 μg/ml anti -CD28 and 1 μg/ml anti -CD3), fewer than 30% of aged T cells expressed the early activation marker CD69, compared to over 70% of young T cells (supplemental figure 1C). Increasing anti-CD28 concentrations slightly improved transduction efficiency (supplemental figure 1D), whereas increasing anti-CD3 had no effect (supplemental figure 1E). Thus, t o improve CAR T cell production yield from aged cells , we employed a "strong activation" protocol (4 μg/ml anti -CD28 and 1 μg/ml anti -CD3), which significantly increased CD69 expression ( supplemental figure 1F), promoted activation -induced cell (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint growth ( figure 1D), and improved transduction efficiency in both young and aged T cells (figures 1E, F). However, CAR T cells gener ated under this stronger activation were functionally inferior to those generated using the weak activation protocol (figure 1G). These findings prompted us to explore alternative strategies to improve CAR T cell production from aged donors. Our previous studies demonstrated that T cells from aged mice fail to increase labile iron pools upon activation, and iron supplementation during activation enhances their proliferation21. Based on these findings, we tested whether iron supplementation during the initial activation step would improve aged CAR T cell s production yield. GFP + CAR T cell analysis revealed enhanced proliferation, as assessed by CellTrace Violet dye dilution, in CAR T cells generated using iron supplementation (figure 1H). This led to a significant increase in the yield of aged CAR T cells ( figure 1I). This optimized protocol enabled the generation of sufficient numbers of aged CAR T cells to analyze their cellular composition and function in comparison to those from young donors. The cellular composition of CAR T cells is affected by donor age CAR T cells are generated from bulk CD3+ T cells. During aging, T cell composition changes are manifested by a shrinkage of the naïve T cell compartment and accumulation of memory and terminally differentiated T cells1. To determine how donor age impacts CAR T cell phenotype, T cells derived from young and aged donors were analyzed at different steps during CAR T cell production (figure 2A). As expected, T cells isolated from young donors were predominantly CD4+ cells, with a naïve phenotype (CD62L+CD44lo; figure 2B), whereas the majority of T cells derived from aged donors were effector memory (EM) cells (CD62L-CD44hi; Figure 2C). After initial activation, young T cells a cquired a central memory (CM) phenotype, showing an (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint increase in the proportion of CD8 + T cells (figure 2D). Aged T cells maintained their EM phenotype and, like young T cells, exhibited an increased proportion of CD8+ cells (figure 2E). Successfully transduced, GFP+ CAR T from young donors were primarily CD8+ cells, with a CM phenotype (figure 2F). However, in T cells derived from aged donors, the majority of successfully transduced cells exhibited a CD4+ EM phenotype (figure 2G). Together, these data demonstrate that the final composition of CAR T cells is influenced by donor age, raising the important question of whether and how this might impact their functionality. T cells from aged donors non-specifically kill target cancer cells independently of CAR expression Upon binding to its target, CD19, the CAR triggers downstream signaling cascades that drive T cell activation, proliferation, and effector functions. Activated CAR T cells eliminate their targets through the secretion of cytotoxic granules containing perforin and GzB, proinflammatory cytokines (e.g., IFN g), and chemokines 22. To compare the functionality of young and aged CD19-CAR T cells, we monitored their killing efficiency against A20 target cells using live-cell imaging. A20 cells were stained with CellTrace Violet (CTV) and cocultured with either young or aged CAR T cells in media supplemented with anti -Annexin V antibodies to detect apoptosis. Cultures were observ ed for 24 hours, and the fraction of Annexin V+ A20 cells (CTV+) was quantified. Controls included A20 cells alone and A20 cells cocultured with primary young or aged T cells after pre-activation, without transduction. Young CAR T cells efficiently killed their targets within the first hour of coculture, with 75% of A20 cells expressing Annexin V, compared to only 30% in the presence of control young T cells (figure 3A). In contrast, aged CAR T cells exhibited a delayed cytotoxic response, with a significant increase in Annexin V signal observed after 5 hours. Unexpectedly, aged control T (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint cells killed target cells as efficiently as aged CAR T cells, suggesting that their cytotoxicity was independent of CAR expression (figure 3A). To further validate these findings, we performed an additional killing assay at varying target- to-effector ratios, measuring target cell killing after 10 hours, the time point of maximal cytotoxicity (figure 3A). Young CAR T cells demonstrated specific, dose-dependent killing and were significantly more cytotoxic than aged CAR T cells, particularly at higher target -to- effector ratios. Notably, aged control T cells exhibited cytotoxicity comparable to that of aged CAR T cells (figure 3B). Both young and aged CAR T cells responded to target recognition by secreting IFN g. As expected, cultures with higher effector cell concentrations contained higher levels of IFNg. Notably, young CAR T cells secreted more IFNg than aged CAR T cells (figure 3C). GzB secretion by young CAR T cells increased as expected compared to young control cells (p<0.0001, one-way ANOVA; comparing only young control and CAR T cells at 1:3 T:E ratio; figure 3D). However, aged T cells secreted nearly twice as much GzB compared to young T cells, independent of target recognition ( figure 3D). These findings suggest that T cell cytotoxicity in aged cells is at least partially mediated by continuous GzB secretion, rai sing the question of whether aged T cells possess inherent cytotoxicity independent of the ex vivo CAR T cell production process. Aged T cells eliminate target cells through the continuous secretion of GzB and IFNg To assess cytotoxicity mediated by aged T cells, T cells were purified from the spleens of young and aged mice and subjected to a brief chemical stimulation with PMA and ionomycin, in the presence of a Golgi inhibitor (Brefeldin A). A substantial portion of aged T cells (approxiamately 26% of CD4+ T cells and over 80% of CD8+ T cells) produced IFNg, significantly (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint more than the portion of young T cells (figure 4A, B). Moreover, while the percentage of young T cells producing GzB was negligible, aged T cells showed a significant increase in GzB production, particularly among CD8+ T cells (figure 4C,D). Consistent with these findings, about 10% of T cells derived from the spleens of aged mice expressed CD107, a marker for degranulation and secretory activity, compared to only 0.6% of young T cells (figure 4E,F). CD107 expression levels were similar in aged CD4+ and CD8+ T cells (supplemental figure 2A). These findings are consistent with previous studies, suggesting that aged T cells exist in a constant state of degranulation23, which could explain why specific target recognition was not required for cytotoxic activity. To test this hypothesis, CAR T cells and control T cells derived from young and aged donor mice were cultured with A20 target cells using a Transwell setup, where effector and target cells were placed in separate compartments (figure 4G). Aged, but not young, T cells induced target cell death, even in the absence of physical contact (figure 4H). Consistently, CD107 expression was twice as high in aged compared to young CAR T cells (figure 4I). These results suggest that aged T cells kill their targets through continuous degranulation and secretion of cytotoxic cytokines. Young T cells acquire a proinflammatory phenotype when transferred into an aged host. Our previous studies demonstrated that exposure to the hemolytic microenvironment in aged spleens induces multiple aging phenotypes in young T cells, including reduced proliferation, and upregulation of CD3921. To test whether exposure to the aged microenvironment in the spleen contributed to the continuous degranulation and cytotoxicity observed in aged T cells, TdTomato+ T cells were isolated from young, transgenic donors and transferred to young or aged C57Bl/6 wild-type recipients. The mice were sacrificed after 3 weeks, and spleens were collected for analyzing the secretory phenotype of transferred, TdTomato+ T cells (figure 5A). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint Similar to aged CD8+ T cells, young CD8+ T cells isolated from aged spleens exhibited increased IFNg (figure 5B) and GzB (figure 5C) production, along with higher surface expression of CD107(figure 5D), indicating elevated degranulation. Interestingly, young CD4+ T cells also showed increased cytokine production and cytotoxicity after exposure to an aged microenvironment (figure 5E-G), though to a lesser extent than CD8+ T cells. Previous studies identified elevation of GzB secretion and degranulation in aged T cells23. Furthermore, a cytotoxic CD4+ T cell subpopulation was identified that accumulates with aging24,25. Our findings suggest these phenotypes are driven, at least in part, by the aged microenvironment. The efficacy of CAR T cell-mediated immunotherapy is reduced in aged compared to young tumor-bearing mice. To investigate whether host age affects the anti -tumoricidal response of CAR T cells in immune-competent mice, we utilized the syngeneic ZsGreen+ ML21 B cell leukemia model. In this system, B cell leukemia was induced in situ through the infusion of bone marrow cells transduced with multiple oncogenes driven by a tetracycline -responsive promoter15. After confirming that ML21 leukemia cells express CD19, the target antigen for our CAR T cells (supplemental figure 3A), we transfused ML21 cells into young and aged mice pre -treated with cyclophos phamide. Doxycycline was administered in drinking water throughout the experiment to induce transformation ( supplemental figure 3B). Peripheral blood analysis, starting on day 8, showed a gradual accumulation of leukemia cells (ZsGreen +). By day 16, approximately 60% of lymphocytes in both young and aged mice were leukemia cells (supplemental figure 3C). Notably, in this first experiment, two mice did not develop leukemia, suggesting some variability in disease induction (supplemental figure 3C). Based on these data, we selected day 10 for CAR T cell administration in following experiments, as this (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint timepoint represented a linear phase of disease progression with comparable ML21 burdens across individual mice ( supplemental figure 3D). To accurately assess tr eatment efficacy without ongoing disease induction, doxycycline was discontinued at the time of therapy initiation (supplemental figure 3D). While disease progression was more pronounced in mice treated with control T cells, by day 15, we observed a sponta neous decrease in ML21 frequency in both groups ( supplemental figure 3E). To mitigate this spontaneous disease resolution , we modified our protocol to maintain doxycycline administration throughout the experiment ( figure 6A). Using this optimized protocol, young tumor -bearing mice were treated with either control or CAR T cells. ML21 levels in the blood continued to rise until day 15, after which we observed a significant reduction in leukemia burden in CAR T cell-treated mice but not in controls (figure 6B). Thus, despite ongoing doxycycline administration, we successfully established an in vivo protocol to evaluate CD19 CAR T cell therapy in C57Bl/6 immunocompetent mice. To assess the impact of an aged microenvironment on CAR T cell efficacy in vivo, ML21 leukemia was established in six young and ten aged mice. All mice received young CAR T cells on day 10, following the protocol outlined in Figure 6A. Survival was significantly better in young mice treated with CAR T cells, compared to aged CAR T cell -treated mice (figure 6C). Moreover, while young mice exhibited a marked reduction in ML21 levels following CAR T cell treatment, aged tumor-bearing mice did not respond and retained high ML21 burdens (figure 6D). Consistent with the lower efficacy observed in aged mice, and despite transfusing similar numbers of CAR T cells to all mice, quantification of CAR T cells in peripheral blood revealed significantly reduced CAR T cell persistence in aged hosts (figure 6E). Our findings highlight that the age of a tumor -bearing mouse significantly influences the efficacy of CAR T cell immunotherapy at multiple levels. Aging reduces CAR T cell (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint production yield, and alters CAR cell function. While exposure to an aged microenvironment enhances T cell and CAR T cell cytotoxicity, this cytotoxic effect is non-specific, and CAR T cells in aged hosts exhibit reduced persistence and impaired tumor clearance in vivo. This aligns with their altered subpopulation composition, which favors an effector memory phenotype over the central memory phenotype seen in younger CAR T cells —an attribute known to impact persistence26. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint

Aging is associated with a progressive decline in both innate and adaptive immune functions, contributing to increased susceptibility to infections, reduced vaccine efficacy, and a higher risk of malignant and inflammatory diseases27-29. Among immune cell populations, T cells are particularly affected by aging. Thymic involution leads to diminished naïve T cell production, while cumulative cellular and molecular defects impair T cell activatio n, proliferation, and memory formation1. Additionally, age-related changes in antigen presentation and stromal cell composition within the l ymphoid microenvironment further compromise T cell functionality30. Given the essential role of T cells in cancer immunotherapy, this study aimed to elucidate how these age-associated changes impact the efficacy of CAR T cell therapy. Our findings demonstrate that aging influences CAR T cell therapy at multiple stages, starting from production of the transduced cells . T cells derived from aged mice exhibit reduced transduction efficiency, and resulted in an altered CAR T cell product composition. Aged CAR T cells contain a higher proportion of CD4 + T cells with an effector memory phenotype, as opposed to the dominance of the CD8+ central memory phenotype observed in younger CAR T cells. Given that CD8⁺ central memory cells are known for their superior expansion and long- term persistence, while CD4⁺ effector memory cells have limited proliferative capacity and persistence, this age -related change in subpopulation composition may negatively impact CAR T c ell durability and therapeutic efficacy 31,32. Thus, a patient’s immunological age may influence the characteristics of their CAR T cell product and, consequently, the therapy’s effectiveness. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint Previous studies have reported conflicting results regarding the cytotoxicity of young versus aged CAR T cells 11-13. Our studies show that while aged CAR T cells exhibit reduced killing efficacy, they display non specific cytotoxicity, as evidenced by increased degranu lation and GzB secretion independent of target recognition. We further show that this phenotype is not intrinsic to aged T cells but is induced by exposure to the aged systemic milieu. Accordingly, even young T cells aqcuired this phenotype when transferred into aged hosts. Non-specific cytotoxicity could harm healthy tissues, and is in line with reports of elevated risk for cytokine release storm (CRS) in aged patients undergoing CAR T cell therapy. Importantly, the demonstrated differences in outcomes of CAR T cell therapy between young and aged tumor- bearing mice likely stem from a combination of factors, including the accumulation of suppressor cells in aged hosts, and the overall inflammatory milieu that may promote activation-induced cell death (AICD) 27-29. Our previous work identified the hemolytic microenvironment of the aged spleen —characterized by iron and heme accumulation —as a driver of T cell dysfunction, including the upregulation of exhaustion markers and reduced proliferative capacity 21. We further demonstrated that aged T cells exhibit resistance to ferroptosis, an iron- and oxidative stress-induced form of cell death, by limiting iron uptake21. While this adaptation enhances survival in a stressful microenvironment, it comes at the cost of impaired proliferation, which relies on iron -dependent processes such as mitochondrial biogenesis and DNA synthesis 33,34. Thus, the aged microenvironment may impair CAR T cell persistence in aged hosts by diminishing their proliferative capacity. In this study we used a second generation CAR containing a co -stimulatory domain derived from CD28, known to promote an effector memory phenotype. It is possible that other CAR designs could improve CAR T cells function and persistence in the aged milieu. For example (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint incorporating other costimulatory domains like 4 -1BB were shown to enhance T cell persistence35. Other approaches to optimize CAR T cell fitness, like the co-expression of key metabolic enzymes36, the elimination of inhibitory signals 37 could further help improve the long-term efficacy of CAR T cells in older patients. Our findings, if proven relevant to human patients , will have important implications for CAR T cell therapy, particularly in the context of aging and "off -the-shelf" allogeneic CAR T therapies38. Universal donor -derived CAR T cells are an attractive solution to overcome the logistical challenges of autologous CAR manufacturing. However, our results suggest that the efficacy of such treatments may be influenced by the recipient’s immune status and the aged lymphoid microenvironment. Understanding and addressing these age -related barriers will be critical in optimizing CAR T cell therapies for older patients.

The authors thank Prof. Ziv Shulman for providing mouse models . Dr. Shelley Schwarzbaum for her assistance in editing the manuscript. Viktoria Zlobin and Dr. Amit Avrahami from the Technion Preclinical Authority for maintaining our aged mouse colony and help with in vivo studies. Dr. Aviv Lutati and Yousef Mansour from the Technion Life Science and Engineering Infrastructure Center for their help with cell sorting. The research was funded by grants given to N.R.-H. by the Israeli Cancer Association (ICA) and the Israel Cancer Research Fund (ICRF). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint Figure Legends. Figure 1: Suboptimal activation impairs the production of CAR T cells from aged donors. (A) Scheme showing experimental design. CD3 + T cells were isolated from the spleens of young or aged C57Bl/6 mice, activated, and transduced with retroviru ses carrying the CAR plasmid. (B) Quantification of transduction efficiency by the percentage of GFP + CAR T cells (each data point represent a pool of aged (n=5) or young (n=3) mice ). (C) a representative FACS plot. (D) analysis of cell size, and (E -F) transduction efficiency in young versus aged T cells following a “weak” (0.5 μg/ml anti-CD28 and 1 μg/ml anti-CD3) or “strong” (4 μg/ml anti- CD28 and 1 μg/ml anti-CD3) stimulation. (G) A20 cells were labeled with CellTrace Violet (CTV) and co-cultured with young CAR T cells, prepared using either weak or strong activation at varying target -to-effector (T:E) ratios. A20 cell death was assessed using PI staining (data points are technical replicates of T cells pooled from 5 young mice). (H) Representative plot showing proliferation of aged CAR T cells prepared with or without supplementation with ferric ammonium citrate (FAC). (I) aged CAR T cell production yield with or without FAC supplementation (data points are technical replicates of T cells pooled from 6 aged mice). Bar graphs represent mean ± SEM (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001); one -way ANOVA with Tukey’s multiple comparisons test (G), or unpaired Student’s t test (B,I)). Figure 2: The cellular composition of CAR T cells is affected by donor age (A) Experimental scheme. Young (n=3, pooled) and aged (n=5, pooled) T cells were activated, transduced with viruses carrying the CAR plasmid, and analyzed by flow cytometry at different steps of the production process, as indicated by the colored frames. (B,C) analysis of T cells immediately after isolation. (D,E) Analysis of T cells following initial activation, and (F,G) (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint analysis gating on GFP+ CAR T cells. Panel shows representative data from 2 independent experiments. Figure 3: T cells from aged donors kill target cancer cells independently of CAR expression (A) CAR T cells were produced from young (n=3, pooled) and aged (n=5, pooled) mice and cultured with CellTrace Violet-labeled A20 cells, at a 1:1 ratio. A20 killing was assessed over 24 hrs using live cell imaging and A nnexin V staining. (B) Analysis of A20 killing after 10 hr coculture with young and aged CAR T cells at varying T: E ratios. A20 cell killing was evaluated by flow cytometry using Zombie Aqua s taining. Media from th is experiment was used to quantify IFNg (C), and Granzyme B (GzB) (D) by ELISA. Bar graphs represent mean ± SEM. Data points are technical replicates. (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001); one -way ANOVA with Tukey’s multiple comparisons test. Figure 4: Aged T cells eliminate target cells through the continuous secretion of GzB and IFNg. Young and aged T cells were subjected to a brief chemical stimulation with PMA and ionomycin in the presence of a Golgi inhibitor (BrefeldinA) for 5 hours. Production of IFNg (A,B) and GzB (C,D) was quantified by flow cytometry. Each point represents data collected from a single young (n=4) or aged (n=4) mouse. (E,F) Analysis of CD107 expression on freshly isolated young and aged T cells. Each point represents data collected from a single young (n=5) or aged (n=5) mouse (G) Schematic showing the setup of a Transwell co-culture assay. (H) Quantitation of A20 killing by young and aged CAR T cells using the Transwell co -culture system. T cells used for CAR T cell generation were pooled from young (n=3) and aged (n=5) mice. Single dot s represent technical replicates. (I) Representative plots showing CD107 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint expression in young vs. aged CAR T cells . Bar graphs represent mean ± SEM. (*p<0.05, ****p<0.0001); one-way ANOVA with Tukey’s multiple comparisons test (A,C,H), or unpaired Student’s t test (E). Figure 5: Young T cells acquire a proinflammatory phenotype when transferred into an aged host. (A) Experimental design. Young T cells from transgenic mice constitutively expressing TdTomato were transferred into young or aged C57Bl/6 wild -type recipients. After 3 weeks, recipient mice were sacrificed, and CD3+ T cells were purified from the spleen, and analyzed by flow cytometry, gating on TdTomato + T cells. Production of IFN g and GzB was an alyzed after a short ex vivo stimulation with PMA and ionomycin in the presence of a Golgi inhibitor (BrefeldinA). CD107 expression was quantified on freshly isolated T cells. CD8 + T cells (B-D), and CD4+ T cells (E-G) were analyzed separately. Error bars represent mean ± SEM. Each dot represents data collected from a single young (n= 8) or aged (n= 11) recipient mouse (**p<0.01, ****p<0.0001); Student’s t-test. Figure 6: The efficacy of CAR T cell-mediated immunotherapy is reduced in aged compared to young tumor bearing mice. (A) Experimental design. Young and aged mice were injected intraperitoneally (i.p.) with cyclophosphamide, followed by transfusion with genetically modified bone marrow cells . Doxycycline was administered in drinking water to induce transformation. On day 10 , the mice received an i.v. injection of either control or CAR T cells. Blood samples were collected via tail vein puncture s. (B ) ML21 leukemia was established in 8 young mice , followed by transfusion of either control (n=4) or CAR (n=4) T cells. Percentage of ML21 from total (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint lymphocytes was determined by quantifying ZsGreen expression. (C-E) Young (n=6) and aged (n=10) mice were inoculated with ML21 leukemia, and treated with young CAR T cells . (C) Kaplan-Meier survival curve. (D) Percentage of ML21 from total lymphocytes was determined by quantifying ZsGreen expression. (E) Percentage of CAR T cells of total lymphocytes was determined using the Myc tag incorporated into the CAR. Data points represent mean ± SEM. (*p<0.05, **p<0.01 ); two-way ANOVA with Tukey’s post -hoc analysis (B, D), and unpaired Student’s t-test (E). Supplementary Figure Legends Supplemental Figure 1: (A) Schematic of the PBullet plasmid encoding a CD19 -targeting CAR with the GFP reporter. (B) Analysis of A20 killing after coculture with young control and CAR T cells, at varying T: E ratios. A20 cell killing was evaluated by flow cytometry using PI staining. (C) CD69 expression on CD3⁺ T cells from young and aged mice after we ak activation. (D -E) CAR transduction efficiency in young and aged T cells stimulated with varying doses of anti -CD28 (D) or anti - CD3 (E). (F) Analysis of CD69 expression on CD3⁺ T cells from young and aged mice following strong activation. Bar graphs repr esent mean ± SEM. Data are pooled from 3 young mice; each dot represents a technical replicate (****p<0.0001); one-way ANOVA with Tukey’s multiple comparisons test. Supplemental Figure 2: (A) CD107 expression on freshly isolated young (n=5) and aged (n=5) T cells. Analyzed by flow cytometry. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint Supplemental Figure 3: (A) CD19 expression on ZsGreen⁺ ML21 cells analyzed by flow cytometry. (B) Experimental design. Young and aged mice were injected intraperitoneally (i.p.) with cyclophosphamide, followed by i.v. transfusion of ML21 leukemia cells. Doxycycline was administered in drinking water. Blood samples were collected via tail vein punctures. (C) ML21 leukemia was established in young (n=5) and aged (n=7) mice. ML21 percentage out of total lymphocytes was determined by quantifying ZsGreen expression. (D) Experimental design. ML21 leukemia was established in 10 young mice, followed by transfusion of either control (n=5 recipients) or CAR (n=5 recipients) T cells. Doxycycline was withdrawn at the time of treatment. (E) Analysis of ML21 percentage of total lymphocytes in the experiment described in D, determined by quantifying ZsGreen expression. (***p<0.001); two-way ANOVA Tukey’s post- hoc analysis. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint

1 Mittelbrunn, M. & Kroemer, G. Hallmarks of T cell aging. Nat Immunol 22, 687-698, doi:10.1038/s41590-021-00927-z (2021). 2 Gross, G., Waks, T. & Eshhar, Z. Expression of immunoglobulin-T-cell receptor chimeric molecules as functional receptors with antibody-type specificity. Proc Natl Acad Sci U S A 86, 10024-10028, doi:10.1073/pnas.86.24.10024 (1989). 3 Cappell, K. M. & Kochenderfer, J. N. Long-term outcomes following CAR T cell therapy: what we know so far. Nat Rev Clin Oncol 20, 359-371, doi:10.1038/s41571- 023-00754-1 (2023). 4 Yin, C. et al. CAR T Cell Exhaustion but Not Ex Vivo Cytotoxicity Is Predictive of Patient Clinical Response: An Interim Analysis of ACIT001/EXC002, a Phase Ib/II Trial of Decentralized Production of CAR T Cells for Treatment of Relapsed/Refractory Aggressive NHL and ALL. Blood 142, 2092-2092, doi:10.1182/blood-2023-187663 (2023). 5 Kouro, T., Himuro, H. & Sasada, T. Exhaustion of CAR T cells: potential causes and solutions. J Transl Med 20, 239, doi:10.1186/s12967-022-03442-3 (2022). 6 Chihara, D. et al. Real-world experience of CAR T-cell therapy in older patients with relapsed/refractory diffuse large B-cell lymphoma. Blood 142, 1047-1055, doi:10.1182/blood.2023020197 (2023). 7 Schuster, S. J. et al. Tisagenlecleucel in Adult Relapsed or Refractory Diffuse Large B-Cell Lymphoma. N Engl J Med 380, 45-56, doi:10.1056/NEJMoa1804980 (2019). 8 Neelapu, S. S. et al. Axicabtagene Ciloleucel CAR T-Cell Therapy in Refractory Large B-Cell Lymphoma. N Engl J Med 377, 2531-2544, doi:10.1056/NEJMoa1707447 (2017). 9 Reyes, K. R. et al. Safety and Efficacy of BCMA CAR-T Cell Therapy in Older Patients With Multiple Myeloma. Transplant Cell Ther 29, 350-355, doi:10.1016/j.jtct.2023.03.012 (2023). 10 Patel, R. D. et al. Clinical Factors Associated with Failure to Manufacture Commercial CAR-T Cell Products Among LBCL Patients. Transplantation and Cellular Therapy, Official Publication of the American Society for Transplantation and Cellular Therapy 30, S214, doi:10.1016/j.jtct.2023.12.278 (2024). 11 Zophel, D. et al. Faster cytotoxicity with age: Increased perforin and granzyme levels in cytotoxic CD8(+) T cells boost cancer cell elimination. Aging Cell 21, e13668, doi:10.1111/acel.13668 (2022). 12 Kotani, H. et al. Aged CAR T Cells Exhibit Enhanced Cytotoxicity and Effector Function but Shorter Persistence and Less Memory-like Phenotypes. Blood 132, 2047-2047, doi:10.1182/blood-2018-99-115351 (2018). 13 Guha, P. et al. Frontline Science: Functionally impaired geriatric CAR-T cells rescued by increased alpha5beta1 integrin expression. J Leukoc Biol 102, 201-208, doi:10.1189/jlb.5HI0716-322RR (2017). 14 Bhaskar, S. T., Dholaria, B., Savani, B. N., Sengsayadeth, S. & Oluwole, O. Overview of approved CAR-T products and utility in clinical practice. Clin Hematol Int 6, 93-99, doi:10.46989/001c.124277 (2024). 15 Keinan, N. et al. Syngeneic leukemia models using lentiviral transgenics. Cell Death Dis 12, 193, doi:10.1038/s41419-021-03477-2 (2021). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint 16 Klebanoff, C. A., Yamamoto, T. N. & Restifo, N. P. Immunotherapy: Treatment of aggressive lymphomas with anti-CD19 CAR T cells. Nat Rev Clin Oncol 11, 685-686, doi:10.1038/nrclinonc.2014.190 (2014). 17 Montagna, E., de Campos, N. S. P., Porto, V. A., da Silva, G. C. P. & Suarez, E. R. CD19 CAR T cells for B cell malignancies: a systematic review and meta-analysis focused on clinical impacts of CAR structural domains, manufacturing conditions, cellular product, doses, patient's age, and tumor types. BMC Cancer 24, 1037, doi:10.1186/s12885-024-12651-6 (2024). 18 Evan, G. I., Lewis, G. K., Ramsay, G. & Bishop, J. M. Isolation of monoclonal antibodies specific for human c-myc proto-oncogene product. Mol Cell Biol 5, 3610- 3616, doi:10.1128/mcb.5.12.3610-3616.1985 (1985). 19 Kurachi, M. et al. Optimized retroviral transduction of mouse T cells for in vivo assessment of gene function. Nat Protoc 12, 1980-1998, doi:10.1038/nprot.2017.083 (2017). 20 Zhang, D. K. Y. et al. Enhancing CAR-T cell functionality in a patient-specific manner. Nat Commun 14, 506, doi:10.1038/s41467-023-36126-7 (2023). 21 Ezuz, D., Ombashe, H., watted, L., Atar, O. & Ron-Harel, N. Haem toxicity in the aged spleen impairs T-cell immunity through iron deprivation. bioRxiv, 2024.2005.2005.592551, doi:10.1101/2024.05.05.592551 (2024). 22 Korell, F., Berger, T. R. & Maus, M. V. Understanding CAR T cell-tumor interactions: Paving the way for successful clinical outcomes. Med 3, 538-564, doi:10.1016/j.medj.2022.05.001 (2022). 23 Jin, J. et al. FOXO1 deficiency impairs proteostasis in aged T cells. Sci Adv 6, eaba1808, doi:10.1126/sciadv.aba1808 (2020). 24 Elyahu, Y. et al. Aging promotes reorganization of the CD4 T cell landscape toward extreme regulatory and effector phenotypes. Sci Adv 5, eaaw8330, doi:10.1126/sciadv.aaw8330 (2019). 25 Elyahu, Y. et al. CD4 T Cells Acquire Cytotoxic Properties to Modulate Cellular Senescence and Aging. bioRxiv, 2024.2001.2014.575313, doi:10.1101/2024.01.14.575313 (2024). 26 Ayala Ceja, M., Khericha, M., Harris, C. M., Puig-Saus, C. & Chen, Y. Y. CAR-T cell manufacturing: Major process parameters and next-generation strategies. J Exp Med 221, doi:10.1084/jem.20230903 (2024). 27 Weyand, C. M. & Goronzy, J. J. Aging of the Immune System. Mechanisms and Therapeutic Targets. Ann Am Thorac Soc 13 Suppl 5, S422-S428, doi:10.1513/AnnalsATS.201602-095AW (2016). 28 Nikolich-Zugich, J. The twilight of immunity: emerging concepts in aging of the immune system. Nat Immunol 19, 10-19, doi:10.1038/s41590-017-0006-x (2018). 29 Quiros-Roldan, E., Sottini, A., Natali, P. G. & Imberti, L. The Impact of Immune System Aging on Infectious Diseases. Microorganisms 12, doi:10.3390/microorganisms12040775 (2024). 30 Lancaster, J. N. Aging of lymphoid stromal architecture impacts immune responses. Semin Immunol 70, 101817, doi:10.1016/j.smim.2023.101817 (2023). 31 Tantalo, D. G. et al. Understanding T cell phenotype for the design of effective chimeric antigen receptor T cell therapies. J Immunother Cancer 9, doi:10.1136/jitc- 2021-002555 (2021). 32 Tao, Z. et al. Impact of T cell characteristics on CAR-T cell therapy in hematological malignancies. Blood Cancer Journal 14, 213, doi:10.1038/s41408-024-01193-6 (2024). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint 33 Ron-Harel, N. et al. Mitochondrial Biogenesis and Proteome Remodeling Promote One-Carbon Metabolism for T Cell Activation. Cell metabolism 24, 104-117, doi:10.1016/j.cmet.2016.06.007 (2016). 34 Frost, J. N. et al. Hepcidin-Mediated Hypoferremia Disrupts Immune Responses to Vaccination and Infection. Med (N Y) 2, 164-179 e112, doi:10.1016/j.medj.2020.10.004 (2021). 35 Guedan, S. et al. Enhancing CAR T cell persistence through ICOS and 4-1BB costimulation. JCI Insight 3, doi:10.1172/jci.insight.96976 (2018). 36 Toledano Zur, R. et al. Genetically engineering glycolysis in T cells increases their antitumor function. J Immunother Cancer 12, doi:10.1136/jitc-2023-008434 (2024). 37 Veliz, K. et al. Deletion of CD38 enhances CD19 chimeric antigen receptor T cell function. Mol Ther Oncol 32, 200819, doi:10.1016/j.omton.2024.200819 (2024). 38 Depil, S., Duchateau, P., Grupp, S. A., Mufti, G. & Poirot, L. 'Off-the-shelf' allogeneic CAR T cells: development and challenges. Nat Rev Drug Discov 19, 185- 199, doi:10.1038/s41573-019-0051-2 (2020). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint A B GFP + T cells (%) GFP CFrequency (normalized to mode) DFrequency (normalized to mode) FSC 31.5% GFP SSC 49.8% 12.4% 21.5% Young Aged weak weakstrong strong Young: weak strong G E FSSC GFP Target:Effector ratio 0 20 40 60 * *** 1:1 1:3 1:5 1:10 A20 Dead cells (%) ** *** 20 40 60 80 H I CTV Frequency (normalized to mode) iron - + CAR T cell yield (*103) (per 3M CD3+ T cells) 0 CD3+ T cells activation transduction expansion young/aged 0 20 40 60 **** young aged young aged young aged weak strong control FAC Figure 1: Suboptimal activation impairs the production of CAR T cells from aged donors. (A) Scheme showing experimental design. CD3+ T cells were isolated from the spleens of young or aged C57Bl/6 mice, activ ated, and transduced with retroviruses carry ing the CAR plasmid. (B) Quantification of transduction effic iency by the percentage of GFP+ CAR T cells (each data point represent a pool of aged (n=5) or young (n=3) mice). ( C) a representative FACS plot. (D) analysis of c ell size, and (E- F) transduction efficiency in young versus aged T cells following a “weak” ( 0.5 μg/ml anti-CD28 and 1 μg/ml anti-CD3) or “s trong” (4 μg/ml anti-CD28 and 1 μg/ml anti-CD3) stim ulation. (G) A20 cells were labeled with CellTrace Violet (CTV) and co-cultured with young CAR T cells, prepared using either weak or strong activation at varying target-to-effector (T:E) ratios. A20 cell death was ass ess ed using PI staining (data points are technical replicates of T cells pooled from 5 young mice). (H) Representative plot showing proliferation of aged CAR T cells prepared with or without supplementation with fer ric ammonium citr ate (FAC). (I) aged CAR T cell production yield with or without FAC supplementation (data points are technical replicates of T cells pooled from 6 aged mice). Bar graphs represent mean ± SEM (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001); one-way ANOVA with Tukey’s multiple comparisons test (G), or unpaired Student’s t test (B,I)). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint B CD4 50.5% CD8 48.8% CD4 63.9% CD8 34.1% A CD4 22.1% CD8 77% CD4 70.5% CD8 27.8% young/aged CD3+ T cells activation transduction expansion CD8 SSC CD44 CD62L CD4 62.6% CD8 37.3% C CD4 73.6% CD8 25.9% CD8 SSC CD44 CD62L D CD8 SSC CD44 CD62L E CD8 SSC CD44 CD62L Young Aged Young Aged F CD8 SSC CD44 CD62L G CD8 SSC CD44 CD62L Young Aged Naive 52.6% CM 19.5% EM 20.8% 7.13% Naive 2.87% CM 19.0% EM 59.4% 18.8% Naive 12.7% CM 51.3% EM 28.6% 7.33% Naive 1.16% CM 15.4% EM 67.5% 15.9% Naive 6.26% CM 85.3% EM 8.09%0.32% Naive 1.81% CM 34.9% EM 62.4%0.85% Figure 2: The cellular composition of CAR T cells is affected by donor age (A) Experimental scheme. Young (n=3, pooled) and aged (n=5, pooled) T cells were activated, transduced with viruses carrying the CAR plasmid, and analyzed by flow cytometry at different steps of the production process, as indicated by the colored frames. (B,C) analysis of T cells immediately after isolation. (D,E) Analysis of T cells following initial activation, and (F,G) analysis gating on GFP+ CAR T cells. Representative data from 2 independent experiments. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint A B 1 10 200 50 100 Time (hours) Zombie + A20 cells(%) 0 10 20 30 40 50 1:1 1:3 1:1 1:3 Annexin V+ A20 cells (%) 5 15 25 150 target:effector ratio young control young CAR T aged control aged CAR T A20 **** **** **** * **** D 0 50 110 90 70 30 15 1:1 1:3 1:1 1:3 target:effector ratio GzB [ng/mL] ** * 0 2 4 6 8 CIfNg [ng/mL] 1:1 1:3 1:1 1:3 target:effector ratio **** **** * **** *** ** Legend: Figure 3: T cells from aged donors kill target cancer cells independently of CAR expression (A) CAR T cells were produced from young (n=3, pooled) and aged (n=5, pooled) mice and cultured with CellTrace Violet-labeled A20 cells, at a 1:1 ratio. A20 killing was assessed over 24 hrs using live cell imaging and Annexin V staining. (B) Analysis of A20 killing after 10 hr coculture with young and aged CAR T cells at varying T: E ratios. A20 cell killing was evaluated by flow cytometry using Zombie Aqua staining. Media from this experiment was used to quantify IFN γ (C), and Granzyme B (GzB) (D) by ELISA. Bar graphs represent mean ± SEM. Data points are technical replicates.(*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001); one-way ANOVA with Tukey’s multiple comparisons test. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint D CGZM B+ T cells (%) 0 5 10 15 CD4 CD8 **** GZM B+ SSC CD107 SSC 10.2% 0.53% CD8 0.57% 1.58% CD4 0 2 4 6 8 10 CD107+ T cells (%) *** Young E G H I 20.1% 39.2% 0 5 10 15 20 * * Zombie + A20 cells CD107 F Young Aged Frequency (normalized to mode) Aged young control young CAR T aged control aged CAR T A20 Young Aged **** **** **** CD4 CD8 A IFNg B 80.6% 20.6% CD8 SSC 26.5% 8.28% CD4 IFNg+ T cells (%) 100 80 60 40 20 0 Young Aged **** **** Young Aged Figure 4: Aged T cells eliminate target cells through the continuous secretion of GzB and IFN . Young and aged T cells were subjected to a brief chemical stimulation with PMA and ionomyc in in the presence of a Golgi inhibitor (Brefeldin A) for 5 hours. Production of IFN (A,B) and Gz B (C,D) was quantified by flow cy tometry . Each point represents data collected from a single young (n=4) or aged (n=4) mouse. (E,F) Analys is of CD107 expression on fres hly is olated young and aged T cells. E ach point represents data collected from a s ingle young (n=5) or aged (n=5) mouse (G) Sc hematic showing the setup of a Tr answell co- culture ass ay. (H) Quantitation of A20 killing by young and aged CAR T cells using the Trans well co-c ulture system . T cells used for CA R T cell generation were pooled from young (n=3) and aged (n=5) mice. Single dots represent technical replicates. (I) Representative plots showing CD107 expression in young vs. aged CAR T cells. Bar graphs represent m ean ± SEM. ( *p<0.05, ****p<0.0001) ; one-way A NOVA with Tukey’s multiple comparisons test ( A,C,H), or unpaired Student’s t test (E) . (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint A E Young Aged 0 10 20 30 40 **** F G B C D 0 10 20 30 40 50 **** Young Aged CD107+CD8+ T cells (%) SSC CD107 young aged CD107+CD4+ T cells (%) SSC CD107 young aged young/aged young T cells (tdTomato+) 3 weeks CD3+ T cells analysis gating on TdTomato+ T cells 0 20 40 60 ** GzB+CD8+ T cells (%) Young Aged SSC GzBGzB young aged 15.7% 0.63% IFNg +CD8+ T cells (%) 0 20 40 60 80 ** SSC IFNg young aged Young Aged 24.3% 39.4% 15.8% 34.9% 0 5 10 15 ** Young Aged GzB+CD4+ T cells (%) SSC GzB young aged 1.7% 4.3% 0 20 40 60 ** Young Aged IFNg +CD4+ T cells (%) SSC IFNg young aged 9.5% 24.3% 8.6% 24.3% recepient recepient recepient recepient recepient recepient Figure 5: Young T cells acquire a proinflammatory phenotype when transferred into an aged host. (A) Experimental des ign. Young T cells from transgenic mic e constitutively express ing TdTomato were transferred into young or aged C57Bl/6 wild- type recipients. After 3 weeks, recipient mice wer e sacrificed, and CD3+ T cells wer e purified from the spleen, and analyz ed by flow cytometry, gating on TdTomato+ T cells. Production of IFN γ and GzB was analyzed after a short ex vivo stimulation with PMA and ionomy cin in the presence of a Golgi inhibitor (Brefeldin A). CD107 express ion was quantified on freshly isolated T cells. CD8+ T cells (B-D), and CD4+ T cells (E-G) were analyzed separately. Error bars represent mean ± SEM. Each dot represents data collected from a single young (n=8) or aged (n=11) recipient mous e (**p<0.01, ****p<0.0001); Student’s t-test. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint A young/aged days1 2 21 cyclophosphamide (ip) ML21 (iv) doxycycline bloodsampling 10 control / CAR T cells(iv) D C ML21 cells in the blood (%) CAR T cells Control T cells ** ** 5 10 15 20 0 20 40 60 80 Days * * Days ML21 cells in the blood (%) 10 15 20 0 20 40 60 80 ** CAR T cells in the blood (%)0 2 4 6 8 * Young Aged 0 10 20 30 0 50 100Probability of Survival B E Days Young recepients Aged recepients Young recepients Aged recepients Figure 6: The efficacy of CAR T cell-mediated immunotherapy is reduced in aged compared to young tumor bearing mice. (A) E xperimental design. Young and aged mice were injected intraperitoneally (i.p.) with cyc lophosphamide, followed by transfusion with genetically modified bone marrow cells. Doxyc ycline was administered in drinking water to induce transform ation. On day 10, the mice received an i.v. injection of either control or CAR T cells. Blood samples were collected via tail vein punctures. (B) ML21 leukemia was established in 8 young mice, followed by transfusion of either control (n=4) or CAR (n=4) T cells. Percentage of ML21 from total lymphocytes was determ ined by quantifying ZsGreen expression. (C-E) Young (n=6) and aged (n=10) mice were inoculated with ML21 leukemia, and treated with young CAR T cells. (C) Kaplan-Meier survival curve. (D) Percentage of ML21 from total lymphocytes was determined by quantifying ZsGreen expression. (E) Percentage of CAR T cells of total lymphocytes was determined using the Myc tag incorporated into the CAR. Data points represent mean ± SEM. (*p<0.05, **p<0.01); two-way ANOVA with Tukey’s post-hoc analysis (B, D), and unpaired Student’s t-test (E). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted May 5, 2025. ; https://doi.org/10.1101/2025.05.04.652106doi: bioRxiv preprint