Treatment with Exogenously Added Catalase Alters CD8 T Cell Memory Differentiation and Function

Cell‐based immunotherapy is a promising approach to cancer treatment. However, the metabolically hostile tumor microenvironment (TME) poses a major barrier to this therapeutic approach. Metabolic reprogramming may enhance T cell effector function and support longevity and persistence within the TME. Metabolic processes lead reactive oxygen species (ROS) production, which are mandatory mediators of signaling and immune cell functions, but detrimental when present in excess. Catalase (CAT) is an intracellular antioxidant enzyme that scavenges hydrogen peroxide (H2O2), a central ROS member with a plethora of biological effects. H2O2 is produced intracellularly and extracellularly, diffusing freely between the two compartments. In this study, it is found that scavenging extracellular H2O2 by CAT supplementation has a major impact on the cell redox state, decreased intracellular ROS, but enhanced activation and altered memory differentiation. Under in vitro chronic activation conditions, CAT treatment favors CD8 T cells with less exhausted phenotype, increased activation and memory markers, and high bioenergetic capacity. Under in vitro acute activation conditions, CAT treatment selectively prevents differentiation transition from the stem cell memory/naive (TSCM/TN)‐ to the central memory (TCM)‐like phenotype, while enhancing activation and polyfunctionality. The study highlights the critical role of H2O2 as a “hidden player” in T cell fitness and memory differentiation.


Introduction
Cell-based therapy is gaining considerable attention as a promising new approach in tumor immunotherapy. [1] However, the metabolically hostile tumor microenvironment (TME) is a major challenge that compromises the success of this therapeutic approach. The TME is deprived of nutrients such as antioxidants impair T cell responses. [27] However, excessive ROS production that disturbs the normal redox state causes toxic effects through the production of peroxides and free radicals that can damage proteins, lipids, and DNA, leading to mitochondrial damage and apoptosis. [18] Hypoxic conditions in the TME result in impaired mitochondrial function and generation of excessive levels of ROS which promote T cell exhaustion. [17] Efficient T cell effector and memory function require a fine balance between mitogenic ROS production and scavenging through antioxidant mechanisms. [20,21,24,27,28] Catalase (CAT) is an antioxidant enzyme that scavenges H 2 O 2 . Previous studies have shown the protective effects of overexpressed CAT against T cell oxidative stress and the benefit for CAR-T cell antitumor responses in vitro. [29,30] NOX is localized at the plasma membrane and produces H 2 O 2 at the extracellular space, and also, excess of intracellular H 2 O 2 can diffuse outside the cells.
In this study, we sought to examine whether scavenging extracellular H 2 O 2 by exogenously added CAT could affect T cell fitness using established in vitro chronic and acute activation models. [16] Under in vitro chronic activation conditions with persistent anti-CD3 activation and interleukin-2 (IL-2), CAT treatment favored CD8 T cells with less exhausted phenotype, increased levels of activation and memory markers, and high glycolytic and mitochondrial capacity. Under acute activation, non exhaustion conditions with IL-2 alone, treatment of CD8 T cells with CAT resulted in enrichment of central memory-like phenotype with enhanced activation, and polyfunctionality. CATtreated cells appeared to possess a large bioenergetic capacity/ reserve available to be released upon activation. Detailed analysis on sorted CD8 effector (T EM ), central (T CM ), and stem cell memory (T SCM )-like cells revealed that the observed memory differentiation imbalance by CAT treatment was caused by selectively preventing differentiation transition from the T SCM -to the T CM -like phenotype, while allowing the cells to become highly activated with an increased functional potential. Further studies will determine if CAT treatment of adoptively transferred antigen-specific T cells will enhance antitumor responses and provide a better understanding of the metabolic and epigenetic features of this unusual T cell phenotype.

CAT Treatment Enhances T Cell Activation State
It has been previously shown that antioxidants that mainly scavenge intracellular ROS such as N-acetyl-cysteine (NAC) impair T cell responses. [27] In addition, scavenging extracellular superoxide did not affect activation, proliferation, and cytokine secretion in TCR-stimulated primary human T cells. [31] However, the effects of scavenging extracellular ROS, in particular H 2 O 2 , on T cell fitness and differentiation have not been studied. In order to first examine the contribution of extracellular H 2 O 2 on T cell activation in vitro, we used exogenously added CAT to enzymatically eliminate H 2 O 2 . We used purified total T cells from mouse splenocytes cultured for 3 days in presence of plate-bound anti-CD3 (5 µg mL −1 ) and anti-CD28 (1 µg mL −1 ) in presence or absence of CAT (0.25 mg mL −1 ≈500 Units mL −1 ) ( Figure 1A). CAT treatment resulted in a significant decrease (≈2-fold) in proliferation ( Figure 1B) and IL-2-mediated cell expansion (≈1.4-fold) ( Figure S1A, Supporting Information), however, viability was comparable between CAT-treated and untreated cells ( Figure 1C,D). In both conditions, cells upregulated activation markers such as CD69 and programmed death-1 (PD-1), but notably CD69 expression was even higher (≈4-fold) in CAT-treated compared to CAT-untreated cells ( Figure 1E,F). PD-1 expression, which is upregulated upon T cell activation, also increased about 2-fold in presence of CAT ( Figure 1G,H). Thus, although CATtreated cells proliferated slower, they displayed a higher activation state. Measurement of H 2 O 2 in the medium confirmed its depletion in the CAT-treated samples ( Figure 1I).

T Cell Activation in Presence of Extracellular CAT Results in Decreased Intracellular ROS
By using the same aCD3/CD28 activation approach of purified T cells as in Figure 1, we examined if exogenously added CAT can affect intracellular ROS. More specifically, we assessed mitochondrial and total cellular ROS by using the fluorescent probes MitoSox and CellRox, respectively. We also assessed mitochondrial mass by MitoGreen (MG) and mitochondrial potential by tetramethylrhodamine, ethyl ester (TMRE) staining. By gating separately on CD4 and CD8 cells, we found that CAT treatment resulted in decreased mitochondrial mass in both subsets, but was particularly significant in CD8 cells (Figure 2A). Mitochondrial membrane potential was not affected in CD4 cells, while for CD8 TMRE was significantly decreased ( Figure 2B). Notably, CAT treatment resulted in significant decrease of both mitochondrial ROS ( Figure 2C) and total cellular ROS ( Figure 2D) in both CD4 and CD8 cells. Thus, elimination of extracellular H 2 O 2 resulted in robust intracellular changes. Surprisingly, although CAT-treated cells displayed a higher activation state in terms of expression of surface activation markers, they had lower mitochondrial mass and ROS production.

T Cell Activation in Presence of CAT Generates CD8 with High Glycolytic and Mitochondrial Capacity
In order to examine the effects of CAT on T cell bioenergetics in more detail, we performed Seahorse MitoStress test using purified CD4 and CD8 T cells. Surprisingly, extracellular acidification rate (ECAR), a marker of glycolytic activity characteristic of activated cells, was very much reduced in CAT-treated CD4 and CD8 cells ( Figure 2E,F). However, CD8 cells distinctly displayed a robust increase in ECAR upon oligomycin and fluorocarbonyl cyanide phenylhydrazone (FCCP) addition which almost reached the levels of the untreated cells ( Figure 2F), indicating a very high glycolytic capacity. CAT-treated CD4 cells had lower oxygen consumption rate (OCR) ( Figure 2G), but CAT-treated CD8 cells displayed a similar OCR pattern with the untreated controls ( Figure 2H). Besides increased glycolytic capacity, CAT-treated CD8 cells also displayed a higher spare respiratory capacity (SRC), which represents the residual mitochondrial energy that can be released under conditions of stress ( Figure 2J). In contrast, CAT-treated CD4 cells had decreased SRC ( Figure 2I). In addition, CAT-treated CD4 cells had very reduced production of adenosine triphosphate (ATP) through mitochondrial OCR ( Figure 2K), while CAT-treated CD8 cells maintained a comparable ATP-linked respiration ( Figure 2L). Collectively, CAT-treatment had distinct effects on the bioenergetics of CD4 and CD8 cells, particularly favoring the generation of CD8 cells with high glycolytic and mitochondrial capacity.

CAT Treatment During In Vitro Chronic T Cell Activation Prevents Exhaustion Phenotype
Our initial findings showed that T cells activated in presence of CAT had significantly upregulated activation marker CD69 ( Figure 1E,F). Since, besides activation, CD69 has also been linked with exhaustion, [32] we employed a recently reported in vitro chronic activation protocol [16] to assess CAT effects on a combination of exhaustion markers such as PD-1, T cell immunoglobulin and mucin domain-containing protein 3 (TIM-3), and Thymocyte selection-associated HMG box (TOX). [17] We stimulated purified CD8 T cells with plate-bound anti-CD3/CD28 antibodies and soluble IL-2 for 2 days followed by another 6 days stimulation with plate-bound anti-CD3+IL-2 and culture passage every 2 days. We had a total of three conditions ( Figure 3A): 1) All steps in absence of CAT (CAT-), 2) Addition of CAT in the first 2 days (CAT+), and 3) Addition of CAT in the first 2 days with also a 2-day interval of CAT treatment during the chronic activation phase (2xCAT+). Chronic stimulation for a total of 8 days without CAT (CAT-) resulted mostly in generation of effector memory (T EM )-like cells with high CD44 but low CD62L expression ( Figure 3B). However, in presence of CAT (CAT+), cells were enriched in a CD44 + CD62L + central memory (T CM )-like phenotype [33] (Figure 3B,C). A second incubation step with CAT (2xCAT+), resulted in most cells being CD44 + CD62L + T CM -like cells ( Figure 3B,C). In addition, chronic stimulation without CAT resulted mostly in generation of PD-1 + TIM-3 + ( Figure 3D,E) and TOX + exhausted cells ( Figure 3F,G). Strikingly, in presence of CAT there was a shift to lower PD-1 + TIM-3 + (Figure 3D,E) and TOX + cells ( Figure 3F,G) which were even more reduced after a second incubation step with CAT ( Figure 3D-G). These results indicated that pretreatment of T cells with CAT to deplete extracellular H 2 O 2 during in vitro chronic activation favored a central memory-like phenotype with potential long-lasting benefit and resistance to exhaustion. . Error bars represent the standard deviation from the mean value. Asterisks indicate P value level of statistical significance (**P < 0.01, ****P < 0.0001).

CAT Treatment During Acute Activation Alters Memory Differentiation and Function by Acting on T SCM Cells
Besides examining the effects of CAT during a chronic activation-exhaustion protocol, we sought to determine the effects of CAT treatment during acute activation in presence of IL-2 alone with the perspective to use this approach in later studies with adoptive transfers. CD8 T cells were first activated with platebound aCD3/CD28 and soluble IL-2 with or without CAT for 2 days and then cells were cultured with IL-2 alone for 6 additional days with cell passage every 2 days. We had a total of three conditions ( Figure 4A): 1) All steps in absence of CAT (CAT-), 2) Addition of CAT in the first 2 days (CAT+), and 3) Addition of CAT in the first 2 days with also a 2-day interval of CAT treatment (2xCAT+). At the end of the course, cells were examined for viability, activation, exhaustion, stemness, and functionality markers. Cells activated in presence of CAT had improved viability which was even more evident after a second step with CAT ( Figure 4B,C). Under conditions of IL-2 without CAT (CAT-), the majority of the cells differentiated into effector memory (CD44 + CD62L -) T EM -like cells ( Figure 4D). In contrast, in presence of CAT (CAT+) there was a robust increase of a central memory-like phenotype (CD44 + CD62L + ) ( Figure 4D,E) which was more increased with the second CAT treatment (2xCAT+) ( Figure 4D,E). Strikingly, the expression of the stemness transcription factor T-cell specific, HMG-box (TCF-1) was doubled in presence of CAT and increased even more with the second CAT step ( Figure 4F,G), indicating that CAT treatment in presence of IL-2 favored a central memory-like phenotype with increased stemness. In order to understand, if CAT treatment was causing a selective expansion or transition to T CM -like cells, we sorted T EM -, T CM -, and T SCM -like cells at day 3 of initial anti-CD3 and anti-CD28-based stimulation ( Figure S2A, Supporting Information). Sorted cells were cultured with soluble IL-2 (10 ng mL −1 ) in presence or absence of CAT (500 U mL −1 ) for another 5 days (Day 8), at which point, cells were analyzed for T EM -, T CM -, and . Error bars represent the standard deviation from the mean value. Asterisks indicate P value level of statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001). E,F) Extracellular acidification rate (ECAR), G,H) oxygen consumption rate (OCR), I,J) spare respiratory capacity (SRC), G,H) ATP-linked OCR after mito-stress test were assessed by Seahorse analysis at 72 h of activation. Blue, yellow, and red arrowheads indicate addition of oligomycin, FCCP, and mixture of Rotenone/Antimycin, respectively. Results are representative of 6 measurements (n = 6). Error bars represent the standard deviation from the mean value. Asterisks indicate P value level of statistical significance (**P < 0.01, ****P < 0.0001).
T SCM -like cell percentages by flow cytometry. Surprisingly, CAT treatment did not selectively expand T CM -like cells, but impaired the transition of T SCM -to T CM -like cells ( Figure S2B, Supporting Information). Since CAT activity in the culture medium progressively decreased (unpublished observations), cells were able to transition to T CM -like phenotype, resulting in a temporarily observed enrichment of T CM -like cells. Notably, this process did not impair, but rather enhanced T cell activation as indicated by an ≈2-3-fold increase of CD69 expression in all T EM -, T CM -, and T SCM -like cells ( Figure S2C, Supporting Information).
In order to understand, if the above changes have any impact on CD8 T cell function, at the end of the stimulation course at day 8, we assessed the ability of the cells to produce interferon gamma (IFN)-γ and tumor necrosis factor alpha (TNF)-α by in vitro stimulation with Phorbol 12-myristate 13-acetate (PMA)/Ionomycin (PMA/I) and intracellular staining. Strikingly, CAT-treatment resulted in higher percentage of TNF-α + IFN-γ + CD8 cells which was even higher upon the second CAT treatment (Figure 5A,B). Notably, the expression level of both cytokines, indicated by mean fluorescent intensities (MFI), was doubled in the CAT-treated cells, and tripled when cells were treated with CAT for a second time ( Figure 5C-E), indicating that CAT treatment not only generated CD8 cells with enhanced viability, memory, and stemness, but also generated cells with enhanced polyfunctionality.
Next, we sought to determine if similar effects can occur on human primary T cells. We stimulated human CD8 cells with anti-CD3 and anti-CD28 for 7 days in absence or presence of CAT (with passaging at Day 3) and then continued culturing them either with 1) IL-2 without CAT (CAT-) or 2) with IL-2 and CAT (CAT+) for another 4 days and cells were analyzed at Day 11 ( Figure 6A). CAT treatment resulted in about twofold decrease in proliferation ( Figure 6B,C). Viability in absence of CAT was about 80%, but in presence of CAT about 60%. The latter was similar to that of unstimulated cells (data not shown), thus, possibly related with the decreased proliferation under CAT treatment. IL-2-mediated cell expansion in presence of CAT was also decreased (≈1.3-fold) ( Figure S1B, Supporting Information). We then went on to examine the activation and memory differentiation characteristics of the generated cells. Human memory subsets are determined based on expression of CD45RO and C-C chemokine receptor type 7 (CCR7). [34] Treatment with IL-2 in absence of CAT (CAT-) resulted in the majority of cells being into an effector memory CD45RO + CCR7 -T EM -like phenotype ( Figure 6D,E). In striking contrast, treatment with IL-2 in presence of CAT (CAT+) resulted in the majority of cells being in a central memory CD45RO + CCR7 + T CM -like phenotype ( Figure 6D,F). Also, treatment with IL-2 and CAT (CAT+) resulted in a small but reproducibly present population of naïve CD45RO -CCR7 + T N -like phenotype ( Figure 6D,G). Although we did not sort individual human CD8 memory subsets, at day 3 of initial anti-CD3 and anti-CD28 activation, human CD8 cells were predominantly at the T CM -like stage (73%) and about 15% at the T N -like stage ( Figure S3A, Supporting Information). Similar to what we observed in mouse cells, after CAT treatment there was a sustained T N population (16.1%) and enrichment of T CM -like cells (59.3%) ( Figure S3B, Supporting Information). The effect of CAT on T N -like cells was more evident when CAT treatment was added at the beginning of the culture, when most cells were at the T N -stage (54.7%) ( Figure S3C, Supporting Information) and were mostly sustained at this stage upon CAT treatment (45%) ( Figure S3D, Supporting Information). In the latter case, there was no T CM -like cell enrichment observed, indicating that the CAT effect was mainly on T N -like cells. Stimulation in absence of CAT resulted in about 1.5-fold increase of CD69 expression in T N -like cells compared to unstimulated cells ( Figure S3E, Supporting Information). Notably, CAT-treatment resulted in even higher (≈3-fold) CD69 upregulation compared to unstimulated cells ( Figure S3E, Supporting Information), indicating that the T N -like cells are not sustained in a naïve state but are efficiently getting activated. Although CAT treatment significantly slowed down proliferation and decreased cell expansion, the cells had significantly higher activation state indicated by multiple markers such as CD25, CD69, and 41BB ( Figure 6F-H). Contrary to what we observed in mouse CD8 cells, in human CD8 cells the expression of the stemness marker TCF-1 decreased ( Figure 6N,O).
Moreover, cultures with IL-2 alone had low expression of the exhaustion markers PD-1, TIM-3, and TOX, all of which were slightly increased upon CAT treatment (Figure 7A-F). Nevertheless, upon PMA/I challenge, CAT-treated cells had significantly higher percentage of IFN-γ + TNF-α + cells ( Figure 7G,H) and also significantly increased expression level of each of these cytokines ( Figure 7I,J). Collectively, similar to what we observed with mouse CD8 T cells, treatment with CAT of human CD8 T cells to deplete extracellular H 2 O 2 during effector differentiation, decreased proliferation and expansion, it sustained memory differentiation at the T SCM /T N stage, eventually resulting in temporary observed enrichment of the T CMlike phenotype. During this process, cells acquired increased markers of activation and polyfunctionality.

Discussion
Recent advances in tumor-antigen specific T cell identification and CAR-T cell development hold great promise for cell-based tumor immunotherapies. However, without new ways to overcome the TME barrier, these approaches will not be efficient. In addition, T cell expansion strategies under continuous stimulation conditions may lead to severe dysfunction and exhaustion. In order to increase the chances of T cell survival in the TME, preactivation regimes should preserve their energy potential to be released at the appropriate time. However, the necessity to reach high T cell numbers makes it challenging to keep the balance between fitness and exhaustion. H 2 O 2 is a cell-permeable molecule, which is produced intracellularly but also extracellularly and can diffuse freely between the two compartments (Figure 8). Excessive H 2 O 2 production during activation can have significant impact on cellular fate and function as it directly affects signaling phosphorylation by inhibiting phosphatases [35][36][37] and can also inhibit enzymes that regulate epigenetic modifications ( Figure 8A). [25] H 2 O 2 can also oxidize lipids in the intracellular or extracellular environment which can have deleterious effects on T cell function. [38] Antioxidant enzymes such as catalase (CAT), glutathione (GSH) peroxidase (GPX), and peroxiredoxin (PRDX) are the main mediators of H 2 O 2 detoxification. CAT, the oldest known and first discovered antioxidant enzyme, is a key enzyme in the metabolism of H 2 O 2 . [39] It mainly resides in peroxisomes and catalyzes the reaction: 2H 2 O 2 → 2H 2 O + ↑O 2, thus besides scavenging hydrogen peroxide it also generates oxygen which serves a unique role as fuel for mitochondrial respiration. [40] In the present study, we examined the effects of exogenously added CAT on T cell activation, fitness, and differentiation to effector and memory phenotypes under either persistent or nonpersistent activation conditions. Commonly used antioxidants such as NAC or GSH significantly impair T cell responses. [27] Indeed, T cell stimulation in presence of CAT significantly decreased proliferation and expansion but surprisingly, resulted in robust increase of activation markers such as CD69 and PD-1 (Figure 1). Although activated cells are usually characterized by increased ROS production which is necessary for activation, [20,21] in this case, elimination of extracellular H 2 O 2 had robust intracellular effects resulting in significantly decreased mitochondrial and total cellular ROS ( Figure 2). This phenomenon may be explained if we consider an osmotically-driven enhanced passive diffusion of excessive intracellular H 2 O 2 to the extracellular space ( Figure 8B). Then, due to limited H 2 O 2 availability, signaling and mitogenic programs required for proliferation are insufficiently activated, but continuous treatment with TCR stimulation or effector induction with IL-2 can still induce enhanced expression of activation markers resulting in accumulation of an enhanced bioenergetic potential and improved polyfunctionality ( Figure 8B). Notably, CAT treatment resulted in decreased mitochondrial potential, particularly on CD8 cells, as indicated by the potentiometric dye TMRE (Figure 2). Low mitochondrial potential and ROS have been shown to be indicative of memory T cell phenotype, enhanced in vivo self-renewal and enhanced antitumor function. [45] Although the CAT-treated T cells had increased activation markers, their bioenergetic profiles for glycolytic activity were surprisingly low at baseline for both CD4 and CD8 cells (Figure 2). CD8 cells, in particular, displayed a robust increase of ECAR upon addition of the mitochondrial ATP synthase inhibitor oligomycin and of the OXPHOS uncoupler FCCP, indicating that CAT-treated cells had enhanced glycolytic capacity. Besides that, CAT-treated CD8 cells also had increased spare respiratory capacity and unaffected OCR-linked ATP production, contrary to CD4 cells in which both of these factors were decreased. These differences suggest that H 2 O 2 may have distinct effects on CD4 and CD8 cells. These unusual and potentially fitness-favorable CD8 characteristics, prompted us to examine whether CAT pre-treatment of CD8 T cells may be beneficial during an in vitro chronic activation course. Indeed, CAT pretreatment of mouse CD8 cells, resulted in lower percentage of cells expressing the exhaustion markers PD-1, TIM3, and TOX at the end of the chronic activation course (Figure 3). Notably, cells were enriched into a central memory T CM -like (CD44 + CD62L + ) phenotype and all these changes were even more pronounced upon a second CAT-treatment during the chronic activation course. This finding may align with a previous observation that H 2 O 2 can preferentially promote cell death of CD8 memory cells. [46] Here, conversely, we found that CAT-mediated deprivation of H 2 O 2 favors the generation of memory-like cells.
We next sought to determine the effects of CAT treatment during acute T cell activation in presence of IL-2 alone as this would be an ideal approach for potentially in latter studies to generate CD8 T cells for adoptive transfers with better fitness and long lasting efficiency. Remarkably, under acute activation conditions, CAT pretreatment resulted in improved viability, in enriched T CM -like (CD44 + CD62L + ) phenotype, enhanced expression of the stemness marker TCF-1, increased polyfunctionality in terms of IFN-γ and TNF-α expression, and these effects were even more pronounced after a second incubation with CAT during the non-persistent activation course (Figures 4 and 5). By treating sorted mouse CD8 T EM -, T CM -, and T SCM -like populations with CAT, we found that differentiation was selectively sustained at the T SCM -like stage ( Figure S2, Supporting Information). Although this seems contradictory to the observed enrichment of the T CM -like cell subset, it can be explained if we consider that CAT activity in the culture medium progressively decreases, and cells are able to transition to the next memory differentiation stages, eventually resulting in a temporary enrichment of T CM -like cells. Similarly, stimulation of human CD8 T cells in presence of CAT sustained their differentiation at the T N -like stage eventually resulting in enrichment of a T CM -like phenotype ( Figure S3C, Supporting Information). As expected, the enrichment of T CM -like cells was not observed when CAT was added from the beginning of the culture, in which case, cells were largely sustained at the T Nstage ( Figure S3D, Supporting Information). Notably, although CAT treatment sustained CD8 memory transition at the T SCM / T N stage, treated cells had robustly increased activation markers and polyfunctionality (Figures 6 and 7; and Figure S2C,E, Supporting Information). Contrary to the mouse CD8 cells, in human CD8 T cells, the expression of the stemness marker TCF-1 decreased upon CAT treatment, indicating that there are several differences in the effects of H 2 O 2 between human and mouse T cells. Overall, these results showed that eliminating H 2 O 2 is a cellular byproduct, a universal molecular player, involved in a myriad of cellular and extracellular biochemical reactions, a "necessary evil" for signaling and epigenetic regulation. [42,44,47] Notably, genetic engineering targeting DNAmethylating enzymes may be a promising strategy for improved cell-based therapies. [48] Here, we observed some potentially beneficial effects of depleting extracellular H 2 O 2 during T cell activation on differentiation to memory, but due to the multifunctional character of the H 2 O 2 mechanism of action, we cannot conclude on a specific mechanism. Since T cell differentiation highly depends on epigenetic regulation, we speculate that the observed effects may be predominantly related with the function of H 2 O 2 -sensitive epigenetic regulators such as DNA methyl-transferases (DNMTs), histone acetyl-transferases (HATs), and deacetylases (HDACs). [43] Our results indicated that in vitro activation of CD8 T cells in presence of CAT selectively prevented the transition of T SCM /T N -to T CM -like cells. As a consequence, proliferation and expansion, features considered important for assessing the success of cell-based therapies, were decreased. However, this effect was not accompanied by impaired activation, but in contrast, activation and functional potential were enhanced.
In addition, it should be noted that the size of the cell product does not always correlate with in vivo expansion or therapeutic success. [49] Besides quantity, the quality of cellular composition and immune phenotype play a crucial role for therapeutic efficacy. [50] Thus, having a much better quality over less quantity of a T cell product may not be detrimental for the final outcome.
In conclusion, our study a) highlights the impact that H 2 O 2 can have as a "hidden player" to T cell fitness and differentiation and b) proposes that elimination of extracellular H 2 O 2 during T cell activation may be a beneficial approach to generate T cells with higher functional potential. Future studies will examine in depth this phenomenon mechanistically and will determine if the in vitro observed properties will prove beneficial when CATtreated cells will be challenged in vivo.

Experimental Section
Mice: All procedures were in accordance with the National Institutes of Health Guidelines for the Care and Use of Animals and approved by the Institutional Animal Care and Use Committee at Beth Israel Deaconess Medical Center (Boston, MA). 8-10 weeks old male or female C57BL/6 (WT) (JAX stock #000664) mice were equally used in the studies.
Cell Cultures: Primary mouse T cells were purified from splenocytes using the mouse T cell (19 851) or mouse CD8 (19 853) negative selection kits (Stemcell Technologies, Cambridge, MA). Purified T cells were Figure 8. A) T cell activation via TCR and CD28 induces generation of ROS, which are required for gene transcription, cytokine production, and differentiation through H 2 O 2 -sensitive transcription factors (TFs) and epigenetic regulators (ERs). [41][42][43][44] TCR activates ROS production through dual oxidase 1 (DUOX1) and NADH oxidase 2 (NOX2), which generate superoxide (O2 •− ) in the intracellular and extracellular space, respectively, which then dismutates to H 2 O 2 either spontaneously or through superoxide dismutase 3 activity. H 2 O 2 can diffuse freely through the plasma membrane. [21] Adding catalase (CAT) exogenously can convert H 2 O 2 to O 2 and it is a question whether CAT-generated O 2 could support mitochondrial function and what impact extracellular H 2 O 2 has on T cell responses, activation, and differentiation. TCR signals also activate protein kinase C family of proteins (PKCs) which lead to ROS production by regulating mitochondrial activity. [21] Through a sequential activation of PI3K and phospholipase A2 (PLA 2 ), CD28 activates 5-lipoxygenase (5-LOX), another important generator of ROS in T cells. [20] B) Activation of CD8 T cells in absence of exogenously added catalase (CAT) leads to production of excessive H 2 O 2 both intracellularly and extracellularly, freely diffusing between the two compartments. These conditions promote differentiation to T EM with decreased metabolic reserves and low polyfunctionality. In contrast, activation in presence of added CAT creates an osmotic imbalance which lowers H 2 O 2 both intracellularly and extracellularly. These conditions sustain cells at the T SCM /T N stage while allowing activation, resulting in increased metabolic reserves and high polyfunctionality.